Skip to main content

Full text of "Amstrad CPC Book: Alan Sugar - The Amstrad Story (1990)(Century)"

See other formats


The Amstrad Story 



!;;:*;;!21! 






as 



Alan Sugar is one of the most extraordinary entrepreneurs to have 
emerged in Britain since the war. He is a man who symbolised many 
of the dramatic social and business developments which transformed 
Britain in the 1980s. 

Sugar has developed Amstrad, which became a publicly quoted 
company little more than ten years ago, into one of the most 
successful consumer electronics concerns in Europe. Unquoted at the 
start of the 1980s, Amstrad regularly doubled its annual profits and 
sales in the middle of the decade, becoming a star of the British 
corporate scene. 

Written with Alan Sugar's co-operation, this book tells the 
remarkable story of how he built up Amstrad from nothing. 
It describes his rise from obscurity in the East End of London to 
become one of the wealthiest men in Britain. 

Time and again, Sugar has revolutionized the high street by 
developing electronic goods that people want at prices they can 
afford. This unbeatable combination came to be known as 'the 
Amstrad effect' and has resulted in Sugar being hailed as a marketing 
genius, excelling in precisely the mass marketing skills where British 
manufacturers have been at their weakest. 

Alan Sugar reveals his business secrets. How he managed to build 
his word processors and computers so cheaply, how Amstrad's 
success in France and Spain highlight the business opportunities in 
Europe in the run-up to 1992, how he learned the lessons of the 
Japanese economic miracle - the Japanese dedication to detail, to 
flexible manufacturing and to mass markets. 

The reasons why Amstrad stumbled at the end of the 1980s are 
also fully described. Sugar was faced with that most difficult of 
business problems: managing Amstrad's transition from its heroic 
period of entrepreneurial hyper-growth to a new phase of corporate 
maturity. The decisions made by Sugar to manage that transition, 
and to position Amstrad for the 1990s, are set out. 

At the heart of this story stands one man: Alan Sugar. Sharp, 
witty, irreverent and unconventional, Sugar's very personal business 
style has brought him into conflict with pillars of the British 
establishment, like the City. It has also ensured that Sugar's singular 
message will stand out in the history of British business. 



£14.99 net 
in UK only 



ALAN SUGAR 



ALAN 
SUGAR 



THE AMSTRAD STORY 



DAVID 
THOMAS 



CENTURY 

LONDON SYDNEY AUCKLAND JOHANNESBURG 



I 



Copyright © David Thomas 1990 

All rights reserved 

Published in Great Britain in 1990 by Century 

An imprint of Random Century Ltd 
20 Vauxhall Bridge Road, London SW1 V 2SA 

Century Hutchinson Australia (Pty) Ltd 
20 Alfred Street, Milsons Point, Sydney, NSW 2061, Australia 

Century Hutchinson New Zealand Ltd 
PO Box 40-086, 32-34 View Road, Glenfieid, Auckland 10, New Zealand 

Century Hutchinson South Africa (Pty) Ltd 
PO Box 337, Bergvlei 2012, South Africa 



David Thomas's right to be identified as the author of this 

work has been asserted by him in accordance with the Copyright, 

Designs and Patents Act, 1988. 

Set in Linotron Baskerville 
Printed and bound in Great Britain by 
Mackays of Chatham, Chatham, Kent 

British Library Cataloguing in Publication data 

Thomas, David 

Alan Sugar: the Amstrad story. 

1. Great Britain. Microcomputer industries. Sugar, Alan 

I. Title 

338.76100416092 

isbn 0-7126-3518-1 



Contents 



1 Sugar Goes Public 1 

2 An Unexpected Child 16 

3 Starting Out 36 

4 The Truck Driver and His Wife 50 

5 A Mug's Eyeful 76 

6 In the Public Eye 93 

7 The People's Computer 115 

8 Birth Pangs in Europe 142 

9 The Amstrad Effect 160 

10 The Boffin and the Entrepreneur 188 

1 1 Sticking to Its Knitting 207 

12 Taking on Big Blue 220 

13 Amstrad Culture 238 

14 Birth of a Multinational 260 

15 Pennies from Heaven 283 

16 Year of Disaster 306 

17 Controlling the Tiger 331 

18 Electric Warrior 346 



Preface 



I became interested in Aian Sugar and Amstrad between 1986 and 
1988, when I wrote about the electronics industry for the Financial 
Times. But it was only after I switched jobs within the FT that I 
approached him with the idea of a book. 

At first he was reluctant. Asked by many writers and publishers 
to tell the story of how he had spearheaded Amstrad's startling 
rise to prominence in the 1980s, his indifference to personal 
publicity had always prompted him to refuse. 

My first thanks must therefore go to Alan Sugar himself for 
finally agreeing to cooperate with the writing of this book. Not 
only did he allow me full access to Amstrad's staff; he also gave 
me introductions to members of his family, including his mother, 
sister and brother, and to friends and early business associates 
who knew him long before he became a figure on the national 
scene. Another journalist once asked Sugar what he did at 
weekends: 'Avoid journalists and photographers' was his prompt, 
heartfelt answer. I realize how much it must have gone against the 
grain for him to be interviewed for this book at length, in his home 
and at the weekend. 

My next debt is to Amstrad's directors and employees who 
willingly agreed to be interviewed for this project. The company 
has a reputation for hostility towards journalists, but this was not 
something which I experienced. 1 must single out for special 
thanks Nick Hewer of Michael Joyce Consultants, Amstrad's 



vn 



PREFACE 



public relations adviser, who has a deep understanding of the 
affairs of his client: only fellow journalists will realize how rare this 
is in his profession. 

I showed Alan Sugar a draft of the book because I was keen to 
remove the inevitable inaccuracies that creep into a story of this 
kind. After receiving his comments, I made some changes to the 
draft. But it was clear to everyone before I started the project that 
it was to be an independent account of the Amstrad chairman and 
his company. I am therefore entirely responsible for the book's 
contents and conclusions. 

I am very grateful to the many people who gave up their time 
to be interviewed for the book. They include Richard Altwasser, 
Ken Ashcroft, Paul Bailey, Matthew Bullock, Richard Clayton, 
Malcolm Cross, Maureen Cross, Greg Delaney, Jose Luis Domin- 
guez, Gerry Eriera, Norma Eriera, David Gibbons, David Gold- 
man, Guy Gordon, John Grant, Graham Grist, Chris Hall, Peter 
Henson, Tim Holland-Bosworth, Peter Hobson, Michael Holling- 
bery, Peter Horrell, Patrick Isherwood, David Johnson, Mark- 
Eric Jones, Stanley Kalms, Diana Kieve, Jeffrey Kieve, Gulu 
Lalvani, Robert Lawrence, Barney Lazarus, Myrna Lazarus, 
Colin Lewin, Nick Lightowler, Ronnie Marks, Gerry Mason, Ted 
Merrette, Malcolm Miller, Greg Morgan, Ashley Morris, Mar- 
garet Mountford, Rupert Murdoch, Howard Myles, Ron Nixon, 
Swraj Paul, Roland Perry, Martin Pluck, William Poel, Vic 
Pollard, Thomas Power, Stan Randall, Jim Rice, Bernhard Schnei- 
der, Sir Clive Sinclair, Peter Smith, Callen So, Fay Sugar, Jimmy 
Tarbuck, David Thomson, Eric Tracey, Marion Vannier, Bob 
Watkins, Lord Weinstock, Arnold Wesker, Terry Wilding, Paul 
Williams, Keith Woolcock, Barry Young and Lord Young. There 
were others who agreed to be interviewed on condition that they 
were not named. 

Some colleagues from the trade press discussed with me aspects 
of Amstrad and the various markets in which it has operated. I 
am particularly grateful to Guy Kewney of PC Dealer, John Lettice 
of Microscope and Arthur Ord-Hume of Independent Electrical Retailer. 
Several magazines allowed me to look through their libraries or 
back copies including Campaign, Electrical Retailer and Trader, Hi-Fi 



• vin * 



PREFACE 

Xews, Microscope, PC Dealer, PC User, Personal Computer World and 
Which Computer?. Brigitte Morel in Dataquest's Paris office shared 
with me her knowledge of personal computer markets in Continen- 
tal Europe, as did Richard Hyman of Verdict his understanding 
of electrical goods retailers in Britain. 

I could not have wished for a better agent than Arthur Good- 
hart, who helped me in ways too numerous to detail. I am grateful 
to Mark Booth, my editor at Century Hutchinson, for his civilized 
approach to his task. My friends Paul Abrahams, Daniel Green, 
Richard Lapper and Raymond Snoddy commented helpfully on 
drafts of some chapters. 

Finally, I would like to thank for everything the two Jennys - 
Jenny Rossiter and Jenny Somerville Thomas - to whom this book 
is dedicated with gratitude, love and affection. 



• ix • 



• 1 • 

Sugar Goes Public 



Pip Greenwell became more and more restless as his chauffeur 
inched his way through the traffic. They had left the City behind 
and swept through what seemed to Greenwell's moneyed eye the 
wastelands of Hackney. Yet as they pushed further north out of 
the centre of London, they had suddenly ploughed into a dense 
line of cars. 

Greenwell was on his way to meet a man called Alan Sugar, 
who was thinking of selling some shares in his audio company, 
Amstrad. It was not likely to be the largest business deal ever 
handled by the senior partner of W. Greenwell, one of the City's 
most prosperous stockbrokers. But Greenwell, a product of Win- 
chester, was a gentleman of the old school. He was punctual to a 
fault and, at this rate, he reckoned they would never reach 
Tottenham on time. He told his chauffeur to get a move on. 

The stockbroker was much relieved when the car finally turned 
into the road housing Amstrad's headquarters, even though the 
sight which greeted him was not one to lift the spirit. Squashed 
between a railway line and a patch of derelict ground, Garman 
Road was completely made over to nondescript medium-sized 
factories and warehouses. Their proprietors were clearly too busy 
to worry about a lick of paint or any other decoration. 

Matters did not improve when Greenwell was ushered through 
a hallway strewn with cardboard boxes into Alan Sugar's presence. 
At 32, Sugar seemed too young to be running a company which 



1 



DAVID THOMAS 

was thinking of going public. As if to underline his youth, the 
Amstrad chairman had parked the knot of his tie a couple of 
inches below the unbuttoned collar of his shirt. He was sporting a 
stubbly beard at a time when a clean-shaven face was part of the 
City's uniform. 

This stocky, bearded youth appeared none too welcoming, as he 
stood there tense and with a scowl on his face. Greenwell wondered 
whether Sugar shared his distaste for being late, and thought it 
politic to ask forgiveness for this sin. T'm so terribly sorry we're 
late, but the fucking traffic was awful.' 

The swear-word sounded odd when uttered in Greenwell's 
cultured voice. There was a moment of silence as it hung in the air 
between the stockbroker and the entrepreneur. Then the tension 
flowed out of Sugar's body and he broke into a grin. 'Thank God 
someone in the City speaks my sort of language,' he said as he 
stepped forward to shake Greenwell's hand. 



New money was meeting old that day in the summer of 1979 when 
Alan Sugar shook Pip Greenwell's hand. It was little more than 
ten years since Sugar had emerged from those same poor Hackney 
streets through which Greenwell had been driven to launch a one- 
man business selling car aerials and other electrical goods from 
the back of a van. He had spent the 1970s criss-crossing Britain 
and the Far East, hunting out ways to undercut his competitors at 
the low end of the audio market. As the decade drew to a close, 
Sugar was about to be rewarded for this period of relentless 
activity. 

The group of the City's best and brightest - who set to work on 
Amstrad's flotation soon after Greenwell's meeting with Sugar - 
had little reason to realize they were dealing with anything out of 
the ordinary. The most optimistic among them hoped that 
Amstrad would consolidate its position in the audio market and 
grow at a respectable but not earth-shattering rate - perhaps 
20-30 per cent a year. Alan Sugar would then be able to rest on 
his laurels, joining the ranks of quietly successful businessmen 



ALAN sugar: the amstrad story 

who are occasionally profiled in the duller reaches of the financial 
press. 

This was not to be. During the next decade, Sugar was to 
become one of the best known businessmen in Britain. Amstrad's 
phenomenal success in the early and mid-1980s, when profits 
regularly doubled each year, sent the value of Sugar's own holding 
in the company into the stratosphere, and by August 1988 his 
fortune stood at £597 million. A few months earlier, he had been 
named by the Sunday Times as the fifteenth wealthiest person in 
Britain. 

The manner in which Sugar led Amstrad was as distinctive as 
the success itself. Amstrad demonstrated a flexibility and a swift- 
ness of foot which was quite out of keeping with the staid and 
stodgy traditions of British business. Just when its potential for 
continuing hyper-growth in ihe audio market was coming to an 
end, the company switched tack and launched into personal 
computers, a graveyard for a whole generation of British 
entrepreneurs. 

Shrugging off a barrage of warnings from City and media 
pundits that he had chosen the worst possible moment to enter the 
computer industry, Sugar spearheaded the formation of the first 
truly mass market for computers in Britain. In the process, he 
affronted received wisdom in the industry by refusing to think of 
computers as pieces of techno-wizardry designed by boffins for 
boffins. Making a reliable computer that was cheap enough for the 
great majority of people to buy and simple enough for them to use: 
that was Sugar's goal. 

This almost banal philosophy succeeded where more sophisti- 
cated attitudes had failed. Amstrad moved computing out of the 
data processing departments into the studies and spare bedrooms 
of Britain, helping to stimulate a revolution in the working habits 
of a whole generation of small business people and other 
professionals. 

In the process, an Amstrad philosophy became discernible, 
elements of which ran through the company's successes in all its 
areas of activity from personal computing, through word process- 
ing, video recorders, music centres, amplifiers and car radios to 



*3 



DAVID THOMAS 



Citizen Band radios and more besides. Amstrad designs cheap 
and simple products by cutting out features which consumers do 
not really want; undercuts its competitors by manufacturing its 
products wherever it finds the best prices; retains the flexibility to 
switch its source of supply whenever a cheaper one emerges; drives 
prices down still further by manufacturing in large volumes for a 
mass market; stimulates that market through massive advertising 
campaigns; and transforms markets by encouraging a whole new 
segment of consumers to start using its simple, inexpensive 
products. 

As the 1980s progressed, much ink was spilt debating this 
formula. Some people argued that it offered little in the way of 
general lessons for British or European business. Although trans- 
parently simple, the Amstrad formula also appeared to be inimit- 
able; it worked on condition that none of the company's 
competitors went down the same route. There was room for only 
one Amstrad in any market, for its philosophy of price cutting 
required it to crowd out its competitors. Indeed, Amstrad often 
began to prepare to quit a market as soon as its competitors tried 
to follow its lead. And, the company's detractors argued, it brought 
little real wealth into Britain because most of its goods were made 
in the Far East. 

Others contended that Sugar had cracked one of the key failings 
of modern British business: the inability to make and market 
consumer products for a mass market. Amstrad's success in France 
and Spain suggested that the company's formula could be 
exported. Ambitious business people eager to understand how to 
compete across Europe in the run-up to the completion of the 
common European market could learn from this unusual British 
company. 

What became known as 'the Amstrad effect' deeply impressed 
some of the hardest and most sceptical businessmen in Britain. By 
the middle of the decade, the General Electric Company- Britain's 
biggest manufacturing group — was considering hiving off its 
consumer division to Amstrad. There was even talk of Alan Sugar 
as a possible successor to Lord Weinstock, G.E.C.'s long-serving 



ALAN sugar: the amstrad story 

managing director - a stunning compliment for a company and a 
man almost unknown a few years earlier. 

Alan Sugar's success began to take on still wider significance. 
He was seen as an archetypal representative of the Thatcher 
decade. Through its deification of competition, Mrs Thatcher's 
Government was shattering institutions long considered to be 
pillars of the British economy: not just those associated with the 
Conservative Party's political opponents, such as the trade unions 
and state-owned industries, but also citadels of British business 
which had failed to adapt quickly enough to the cutting wind of 
international competition. What had appeared to be permanent 
fixtures of the City itself, including W. Greenwell, were to disap- 
pear into financial conglomerates under the pressures of the 'Big 
Bang'. 

A creative destabilizer of markets like Alan Sugar seemed to 
embody the positive message of Mrs Thatcher's economic philos- 
ophy, the other side of the coin to Thatcherism's rooting out of 
dead wood from the undergrowths of the British economy. Brit- 
ain's established electronics companies had become fat and lazy 
on easy profits made by selling over-engineered products to the 
Ministry of Defence and British Telecom, two giant monopoly 
customers. Amstrad chose the harder route of competing head-on 
with the Japanese in the cut-throat arena of Britain's high streets, 
where Sugar was vying for the favour of millions of people who 
had been made affluent by the Thatcher decade - people, in fact, 
just like himself, even though his affluence was on a vastly different 
scale. Alan Sugar and Amstrad were the high priests of Britain's 
new religion, the religion of consumption. 

Sugar seemed to confirm Thatcherism's social vision too. Born 
with no advantages whatsoever, brought up on one of the poorest 
estates in East London, this Cockney entrepreneur now only just 
into his forties had amassed almost unimaginable wealth in barely 
more than a decade. The message rang out from the Amstrad 
story that anyone could make it in Britain's new land of oppor- 
tunity if they were prepared to try their arm and work hard. The 
work ethic Sugar embodied and engendered among Amstrad's 
employees - he joked that he threw a sandwich at them for lunch 



DAVID THOMAS 



as they sat at their desks, if they were lucky - chimed perfectly 
with the Government's message that British workers had to stop 
clock-watching and get on with the job. 

Yet although Sugar drove his workers hard, he also generated 
tremendous loyalty among those who lasted the pace. This loyalty 
was won partly by his distribution of generous share options to 
many of Amstrad's staff, allowing them to share in the company's 
success. More fundamentally, it stemmed from the buzz which 
people experienced when they worked with him - the surge of 
adrenalin which flowed from the feeling that this was a man 
rewriting the rules of the market-place. 

Sugar symbolized the 1980s by refusing to be assimilated into 
the establishment once he had arrived. You would not find him in 
one of London's gentlemen's clubs. He had no time to waste trying 
to shine among the ranks of business leaders in lobbies like the 
Confederation of British Industry. As chairman of one of Britain's 
fastest growing public companies, he continued to speak in exactly 
the same way as when he first left the East End. Funny, aggressive 
and peppered with swear-words, his language stood out from the 
reams of public relations pap uttered by most leaders of corporate 
Britain. 

The Amstrad boss was an icon for popular capitalism, the 
conscious attempt by the Government to teach the mass of British 
people to identify with business in the way that is second nature 
to their American cousins. The use of the privatization programme 
to spread share ownership meant that companies like British 
Telecom became the flagships of popular capitalism. But most 
British people cannot identify with the Oxbridge-educated bureau- 
crats who tend to run these monoliths. By comparison, Sugar was 
a perfect role model, as the Government acknowledged when it 
asked him to spearhead its 1992 awareness campaign. He could 
speak directly to ordinary British people in words they understood. 

Sugar did not alter his lifestyle as he entered the ranks of 
Britain's super-rich; he did not use his position to seek out the 
friendship of the rich and famous. His tastes remained the same as 
those hundreds of thousands of nameless consumers who bought 
his products. Many of his closest friends had similar upbringings 



■6 



ALAN sugar: the amstrad story 

to his own. Sugar shared with them a common journey out of the 
East End of London into the north-east outskirts of the capital, 
differing only in the size of the house he could afford. Like many 
of the most prominent successes of the Thatcher decade - includ- 
ing the Prime Minister herself- he was both part of and curiously 
outside the new order. 

Sugar's dislikes remained those of the outsider and with charac- 
teristic vigour he attacked a broad list of hate figures. The City, 
journalists, boffins, corporate wimps, the established electronics 
companies in Britain and Europe, Americans, the European 
Commission - all were to feel the lash of his tongue. Suspicious of 
people who owe their position to privilege rather than to hard 
work, he is instinctively against anyone whose talents lie simply in 
working the system. He reserves a special contempt for bureau- 
crats, producers of verbiage or regulations which make it more 
difficult for people like himself to earn an honest penny by bringing 
the latest gadget to the masses. 

Many of those attacked by Sugar - and the list included people 
with influence in the corridors of the old Establishment - returned 
the compliment by waiting for his downfall. The decade was 
littered with confident assertions that Amstrad was past its peak 
and with predictions that the company was about to fall flat on its 
face. And as the 1980s drew to a close, the bleakest prophecies of 
Sugar's critics seemed to be coming true. 

Amstrad stumbled as Sugar grappled with a formidable array 
of problems. The growing volume of components it consumed 
made it newly vulnerable to the roller-coaster swings from feast to 
famine and back again in the worldwide supply of electronic parts. 
The company was forced to modify its manufacturing philosophy, 
as trade pressures mounted against the flood of cheap imports 
from the Far East and workers in countries like Taiwan began to 
agitate for improved living standards. Amstrad's tradition of 
running the tightest of ships deprived it of the engineering 
resources needed for the smooth management of its move into a 
more complex line of business computers. It had to struggle hard 
to overcome its cheap-and-cheerful image when promoting higher 
value computers to large business customers. 



DAVID THOMAS 

Most of Sugar's wealth was tied up in the value of his stake in 
Amstrad. Since he owned about 250 million Amstrad shares, every 
Ip change in the share price was worth £2.5 million to him. The 
company's stock came under severe pressure in the late 1980s, 
when general stock market nervousness coincided with the emerg- 
ence of difficulties at Amstrad. By September 1989, the paper 
value of Sugar's fortune had been slashed by £400 million, 
reducing him to his last £197 million and making him the largest 
stock market 'loser' of all time, according to one calculation. 

The root question facing Sugar as he prepared Amstrad for the 
1990s was this: would he be able to manage the company's 
transition from its heroic entrepreneurial phase of hyper-growth 
into a more mature future in which Amstrad would take on the 
settled features of a large, multinational corporation? Put more 
simply still: are Sugar and Amstrad destined to be a flash in the 
pan, like other heroes of British business before them, or are they 
set to become part of the permanent landscape of the British 
economy? 

Sugar ended the 1980s by showing that he understood he was 
facing one of the most difficult of business problems. He had to 
transform the managerial culture at Amstrad from that of a small 
business to one better suited to its new status as a large corporation 
- without destroying the entrepreneurial drive which powered its 
original growth. He strengthened the senior management line-up 
at Amstrad, built up the company's in-house research and devel- 
opment capacity and moved some of its manufacturing back to 
Europe. In short, Alan Sugar began to lay the foundations for a 
new phase in the Amstrad story. 



The team of professionals put together in 1979 to handle Amstrad's 
flotation was preoccupied with a pressing set of questions. They 
needed to overcome the doubts raised by potential investors when 
they were first told about the company. 

What exactly were the big investing institutions, like the insur- 
ance companies and pension funds, being asked to buy into? 
Amstrad had neither assets nor staff to boast about. Its Tottenham 



•8 



ALAN sugar: the amstrad story 

headquarters in North London would not raise much if ever it 
came to the crunch; although the company's main physical asset, 
it was valued at under £400,000. Amstrad's net tangible assets at 
the time totalled £3.7 million, significantly less than the £8 million 
price tag the stock market was being asked to put on the company. 
Neither could Amstrad point to a great array of talent among its 
staff: it had just 117 employees and only a handful of senior 
managers of any calibre in whom investors could have confidence. 

The answer to the institutions' question was, in fact, simple. 
Investors were being asked to back the track record and potential 
of one man, Alan Sugar. But although to the point, this answer 
made some investors nervous. 

Most basically of all, there was the falling-under-a-bus worry; 
in 1979, Amstrad lacked anyone to replace Sugar if he suddenly 
disappeared from the scene. Secondly, there was a question mark 
about Sugar himself. True, he had already notched up a decade of 
impeccable results, but he was still only just into his thirties. How 
could institutional investors be assured that this young man was 
not about to run out of steam? The flotation committee assembled 
by Kleinwort Benson - Amstrad's merchant bank and main 
financial advisers - spent much time discussing how to overcome 
negative reactions to Sugar's youth. 

Third, there was the problem posed by Amstrad's area of 
operations, the audio industry. The roll-call of established British 
audio companies which had come to grief was long and ominous. 
Not only did the audio business demand almost continuous 
innovation embodied in a never-ending stream of new products, it 
was also the field of combat chosen by some of Japan's biggest 
conglomerates. Tt was a very competitive market, dominated by 
some massive companies, and people just wondered whether there 
was ever going to be any profit in it over the years,' remembers 
Tim Holland-Bosworth, a Kleinwort Benson director assigned to 
the Amstrad flotation. 

These concerns were reflected in the smattering of press cover- 
age which Amstrad received before the flotation - the start of a 
nervy love-hate relationship between Sugar and the media. 

The Financial Times'^ influential Lex column pointed out that 



DAVID THOMAS 

'the history of the U.K. hi-fi industry is littered with still-twitching 
casualties', though Lex hedged its bets by praising Amstrad's 
flexibility and profit record. The Financial Weekly posed a question 
which was to recur with monotonous regularity down the years: 
'Will it be the same perky animal in two to three years' time when 
chips are ten a penny?' The Daily Telegraph mused delicately about 
Sugar's health: 'Dynamism in the company springs from Sugar 
himself, and one wonders what might become of Amstrad should 
he, for instance, lose interest or be unable to continue for any 
reason.' 

Kleinwort Benson and the rest of the flotation team pointed to 
Amstrad's record to quell these doubts. Ten years of excellent 
results had left the company with no debt, a return on capital 
above 30 per cent in the second half of the 1970s and pre-tax profit 
margins comfortably above 12 per cent during most of the decade. 
These results were ample testimony to the character of the man in 
the driving seat. 

After a decade of running all aspects of his own show, what 
Sugar did not know about the audio market was not worth 
knowing. He entertained his City advisers by explaining the 
importance of the 'dark hours', the autumn evenings in the run-up 
to Christmas when working-class consumers bought car radios, 
cassette players and all the other items in Amstrad's product line. 
Sugar was also bubbling with ideas for new products. 'The 
impression you came away with was of someone who knew r exactly 
what he was speaking about in his own field. He knew what he 
was doing. The contrast with the average managing director in 
British industry was quite marked,' recalls Howard Myles, then in 
GreenwelPs corporate finance department. 

Most impressive of all from a City perspective was Sugar's 
attitude to profits. 'He had a major obsession with margins. That 
was the thing which drove him. When looking at a new product, 
he always worked back from what he considered an appropriate 
margin,' Holland-Bosworth says. 

Obsessed with profits, Sugar talked the language of business. 
His concerns were with cost, delivery and reliability. This was talk 
which the City could understand much more readily than the 



10 



ALAN sugar: the amstrad story 

techno-babble favoured by some electronics entrepreneurs. 'He 
spoke more familiar language to me than lots of other people in 
technical areas. We have always been nervous of companies 
trading in technology that we don't understand,' notes Holland- 
Bosworth. 

Crucially, some important investors agreed with the positive 
message put out by Amstrad's own advisers. Paul Williams, an 
investment manager with Clerical Medical & General Life Assur- 
ance, was initially worried about the risks inherent in Amstrad's 
line of business, so he went to an investor presentation given by 
Sugar and came away impressed by the philosophy of total 
flexibility articulated by the Amstrad chairman. 

'What attracted us to the company at the time of the flotation 
was that they were very flexible in their sourcing. They would 
source from the cheapest place in the world, wherever that 
happened to be, and then undercut the competition,' explains 
Williams, who persuaded Clerical Medical to become the largest 
single investor in Amstrad after Sugar himself. 

Nevertheless, it was an uphill struggle convincing some insti- 
tutional investors to overcome their fears. The more conservative 
institutions were worried by their inability to compare Amstrad's 
performance with that of another outfit, for as Tim Holland- 
Bosworth admitted at the time: 'We have failed to find a U.K. 
company even broadly similar.' 

In the event, many of the largest institutions refused to buy 
Amstrad's shares, leaving individual investors to play a crucial 
role in securing the eventual success of Amstrad's debut on the 
Stock Exchange. But the Amstrad boss and his team of advisers 
had to put in a lot of work before they reached that stage. 

Sugar was later to tell some of his best jokes and coin some of 
his most colourful invective about his first contacts with the City. 
His account of how he first came across p/e ratios - the ratio 
between a share's price and its earnings, and one of the City's 
favourite tools for measuring company performance - entered the 
Sugar mythology as soon as he related it in a public lecture in 
1987: 



11 



DAVID THOMAS 

The first time I came to the City was in 1979 when I met up 
with Kleinwort Benson. I sat in a large board-room with a lot of 
people who spoke a strange jargon - cash flow, p/e forecasts etc. 
The last time I heard the expression p/e, I was at secondary 
school and it meant physical education. 

So my mind drifted, oblivious to this strange mumble of 
exchange of views from all the so-called advisers around the 
table. I was suddenly woken with the question, fc Mr Sugar, what 
p/e do you think you have?' 

I replied, Twenty press-ups every morning.' 

The funny thing is that they weren't shocked and went on to 
talk about press-ups. 

Such remarks have created the impression of a character aggress- 
ively hostile to the City and all its works. But the Amstrad 
chairman made quite a different impression on those City people 
he worked with during the build-up to the flotation. Huge differ- 
ences in background and language had to be overcome. Sugar's 
drafted statements for the prospectus committee read exactly as 
he spoke: they reproduced his grammar and vocabulary precisely, 
minus only the swear- words. The City team had the unusual job 
of translating his rich Cockney diction into language acceptable to 
the money men. Yet, despite the difference in cultures, Sugar 
buckled down to this task without any fuss. His advisers found 
him to be a rigorously practical young businessman when dealing 
with the issue at hand: screwing the best possible deal out of the 
City once he had decided to sell a quarter of his company, the 
minimum allowed under Stock Exchange rules. 

This pragmatic professionalism shone through in Sugar's 
approach to the flotation. Guided to Kleinwort Benson by Michael 
Hollingbery - chairman of Amstrad's most important customer, 
the Comet retail group - Sugar accepted Kleinwort's advice to 
assemble a flotation team consisting of some of the most prestigious 
of City names. Besides Kleinwort themselves, Amstrad hired W. 
Greenwell as brokers, Touche Ross as auditors and Herbert Smith 
for its legal expertise. 

The Amstrad boss later explained his choice of this quartet of 



12 



ALAN sugar: the amstrad story 

advisers: Tm a great believer in professionals. If someone says to 
me that these people are experts in financial matters, I believe it. 
It's as though I'd had a pain in my chest, well, I'd have gone and 
seen a doctor in Harley Street.' 

The City professionals Sugar had to deal with were mostly of a 
class and background which he had never encountered before. 
Kleinwort Benson's Tim Holland-Bosworth is the very model of a 
double-barrelled City gent: favouring the widest of stripes in his 
pin-striped suits, there is not a hair or a syllable out of place in his 
carefully combed blond locks or his cut-glass accent. Having 
clawed his way up from the East End, it would have been only 
natural for Sugar to have bridled at this epitome of City privilege. 
But he did not. He had no hang-ups about dealing with people 
like Holland-Bosworth, whom he retains as a financial adviser to 
this day. 

True, some of those working with Sugar detected in him an 
apprehensive air at the start of the flotation process, fuelled by a 
suspicion that the City's whizz-kids might be about to rip him off 
- not surprisingly, since Amstrad's relatively simple flotation cost 
a cool £183,000. In private meetings with his advisers, Sugar 
insisted that the meaning of every syllable of City jargon was spelt 
out in words he could understand. He was a quick learner, 
particularly where anything touching the value of money was 
concerned. 'If we were talking about measuring performance by 
something called the p/e ratio, he'd find out pretty damn quick 
how much that meant in his pocket at the end of the day,' 
Holland-Bosworth recalls. 

In public presentations to potential investors, there were times 
when Sugar compensated for his lack of familiarity with City ways 
by being a little brash, a little too simple and aggressive in his 
replies. This manner helped to disguise a certain shyness in public, 
which only those closest to him detected and which it was to take 
him years of public exposure to overcome. 

But there was no doubt that long meetings in plush City offices 
were not Sugar's idea of time well spent. He delegated as much of 
the detailed work as possible to Jim Rice, Amstrad's company 
secretary, making it plain that he would rather be on the road 



13 



DAVID THOMAS 



selling to customers. Where his presence was unavoidable, he held 
as many meetings as possible on home territory. Sugar made his 
pitch to Clerical Medical's Paul Williams in a side-room during a 
break from a trade fair; Williams was used to somewhat more 
formal settings for investor meetings. Holland-Bosworth often had 
to take his team on the unfamiliar train journey across Hackney 
Marshes to Tottenham, timing their travel so as to avoid Totten- 
ham Hotspur fans heading for an evening match. Occasionally 
their calculations went awry and the well-dressed team from the 
City spent an uncomfortable journey eyeball-to-eyeball with a 
trainload of Spurs supporters. 

Once they reached the safe haven of Amstrad's Garman Road 
headquarters, they found a set of offices somewhat more austere 
than their own. Tt was obvious that Alan was not likely to go in 
for atriums or Henry Moore sculptures,' Holland-Bosworth 
remarked with well-bred understatement. When David Thomson, 
Greenwell's head of corporate finance, casually commented on 
being served coffee out of a machine in the chairman's office, 
Sugar put him straight: 'We're not having bloody china in this 
office.' 

The flotation team had more important things to worry about 
than the quality of the coffee cups. As 1979 turned into 1980, the 
British economy headed into the deep recession that marked the 
onset of the economic policies introduced by the first Thatcher 
Government. This caused considerable nervousness about the 
implications for Amstrad, a company dependent on consumer 
prosperity. At the end of January, with the flotation date fast 
approaching, the team was too busy to give more than a passing 
glance to news which was later to have considerable implications 
for Amstrad; Give Sinclair, the electronics entrepreneur, helped 
to trigger the home computer revolution in Europe by launching 
the world's smallest and cheapest computer at an exhibition in 
Wembley. 

In the late 1970s, flotations were less glitzy occasions than they 
are today. Sugar's advisers did much of the selling of the Amstrad 
issue over the phone, trading on their contacts. They were 
confident that one or two institutions like Clerical Medical would 



. 14 



ALAN sugar: the amstrad story 

take large stakes. A week before the 23 April flotation date, 
Kleinwort Benson was able to announce that 5,800 applications 
had been received for a total of 23 million shares - over-subscribing 
the offer nine times. The only disappointment for investors was 
that the Amstrad stock did not open at a slightly higher premium: 
on 23 April 1980, its first day of trading, the share opened at 92p, 
an 8 per cent premium on the 85p offer price. 

This scarcely bothered Sugar. What he cared about was the 
successful sale of all the 2,330,000 shares on offer in Amstrad. He 
personally raised almost £2 million by parting with a quarter of 
his company, while also still holding three-quarters of the shares 
in a publicly quoted company valued at £8 million. Another East 
End millionaire had been born. 

Alan Sugar banked his cheque and returned to what he loved 
best: raking up profits by selling audio equipment. But by floating 
Amstrad on the Stock Exchange, he would never be able to forget 
the City again. Guy Gordon, who had been Sugar's accountant 
when he first set out in business, phoned after the flotation to 
congratulate him and to offer help if he should ever need it again. 
'Guy, I've got so many advisers, lawyers, accountants, round me, 
they're driving me mad. I don't know what half of them do,' was 
the reply. 

This did not stop Sugar from throwing a big party to celebrate 
Amstrad's debut as a public company at the White Elephant, a 
favourite night spot with showbiz personalities on the banks of the 
Thames in central London. Over a hundred people were there, 
including the flotation team with their partners and Sugar's friends 
and family. It was one of the few occasions when his two worlds, 
public and private, ever came together. 

Sugar's parents, Fay and Nathan, stood around, simultaneously 
proud of and bemused by their son's success. Shortly before, he 
had shown his father evidence of the fruits of the flotation. It was 
a significant rite of passage, as Sugar explains: 'It wasn't until I 
physically put a cheque for two million pounds under his nose that 
he realized it was O.K. for him to stop worrying, because I 
actually could make a living without working for somebody.' 



• 15 



An Unexpected Child 



No one had expected Alan Sugar. The entire family was surprised 
when Fay Sugar announced late in the summer of 1946 that she 
was expecting her fourth child. Eleven years had passed since she 
had last given birth and, at 38, she had every reason to believe 
that her child-rearing days were over. 

The news was a mixed blessing for the Sugar family. Fay, a 
bright and warm woman, was naturally delighted at the prospect 
of an addition to the family. Yet there was no escaping the difficult 
circumstances into which the baby would be born. Fay and 
Nathan Sugar were only just picking up the pieces after the war. 

Their move to Clapton, in the East London borough of Hack- 
ney, had been a step in the right direction. It was one remove from 
their old home in Stepney, in the heart of the Jewish East End, 
where families were squashed together in terrace houses lacking 
the most basic facilities such as bathrooms even on the outbreak 
of the Second World War. Clapton was no less working class, but 
the local council had transformed the area shortly before the war 
by building new blocks of council flats. Just as important, Clapton 
was one step nearer the respectability of London's northern 
suburbs. 

Yet no sooner were Fay and Nathan installed in their new 
council flat at the top of a block of flats called Woolmer House on 
the Northwold estate, than the German bombers began to pound 
Hackney into the dust. None of the Sugars' three elder children 



16 



ALAN sugar: the amstrad story 

spent all their childhood with their parents; they were dispersed to 
different corners of England in the evacuation, leaving their father 
and mother to scurry across the playground beneath their flat to 
the air-raid shelter to avoid the bombs. 

The end of the war did not bring much immediate relief. Basic 
foodstuffs were still in short supply. Hackney council could not lay 
its hands on enough slates, cement and plaster board to repair the 
local devastation. With another child on the way, the Sugars' 
three-bedroomed flat suddenly did not appear as spacious as when 
they first arrived. 

In the event, the birth was not the easiest and Fay had to spend 
three weeks in Hackney Hospital, where the doctors insisted on a 
Caesarean birth in view of her age. It was thus with some difficulty 
that, on 24 March 1947, Alan Michael Sugar emerged into the 
world. Remembering the hospital, Fay Sugar - now into her 
eighties but still sharp - says 'her Alan' showed early signs of his 
later character. 'Whilst he was there, he always seemed to be like 
a boss overriding everyone. He made such a lot of noise.' 

Yet as he was growing up, Alan Sugar struck people as a quiet, 
almost solitary child, though there were plenty of other children 
on the estate if he had wanted to surround himself with playmates. 
For a start, he had a brother and two sisters: Derek and Daphne, 
the twins, and Shirley, the eldest of the Sugar children. As a result, 
Alan had to share a bedroom - sleeping in a narrow bed on the far 
side of the room - until his sisters left home to get married. 

The 23 blocks of flats which make up the Northwold estate were 
full of families with young children like the Sugars. Northwold was 
typical London council housing of the era before the capital was 
disfigured by twenty-storey tower blocks. Most of the blocks on 
the estate were like Woolmer House, four storeys tall with half-a- 
dozen flats on each landing divided into two sections by a central 
staircase. The narrow corridors running along each landing have 
acted as playgrounds to generations of children, as have the 
courtyards and patches of grass below. 

Yet Derek, Alan's elder brother, cannot remember ever looking 
over the balcony and seeing him kicking around a football between 
the flats, as he and his friends had done. Brenda, Derek's future 



17 



DAVID THOMAS 

wife, thought Alan was living an unusually dull life for a young 
boy when she started visiting the Sugars. She once prodded her 
younger brother into taking Alan out with his pals, but the boy 
seemed to resent it. As he was growing up, the last thing that Alan 
Sugar appeared to be was an extrovert or a leader of the pack. 

Alan's upbringing was close to that of an only child. The twins, 
Derek and Daphne, were already aged 12 when he was born and 
were too old to act as real companions. Diana Kieve, a friend of 
the Sugars since before the war, remembers Alan as a toddler 
trying to follow Derek around the estate. A teenager with his own 
gang of friends, Derek paid scant attention to his young brother; 
by the time he had done his stint in the Army, he had drifted 
apart from Alan. The youngest Sugar's brother and sisters seemed 
to be of another generation: 'It was like having more adults 
around, growing up in an adult environment really,' he recalls. 

It fell to Daphne, the younger sister, to spend most time with 
the baby of the family. Looking back, she sees early signs of her 
brother's later restless character in what happened on his first day 
at school. She took the five-year-old Alan to his new school, 
Northwold, a severe stone-built primary about a quarter of a mile 
from Woolmer House. It was with some surprise that she opened 
the door of their flat at 1 1 o'clock to find him there, none too 
pleased to be told he had mistaken the morning break for the end 
of the school day. 

Alan found himself growing up in a family which had always 
struggled for money. Nathan Sugar was a semi-skilled worker in 
the garment trade. He tacked up the clothes, preparing them for 
the master machinists to sew together. The money was not bad, 
compared with some of the East End trades, when there was work 
to be had. The trouble lay in the insecurity of the job. 

In the inter-war years, tailoring had been badly hit by the 
depression and the growing mechanization of the small garment 
factories. Even when the country returned to prosperity in the 
1950s, the work remained seasonal and Nathan would be laid off 
without pay at a moment's notice, sometimes for weeks at a time. 
In his earliest years, Alan did not fully grasp the pall of insecurity 
thrown over the whole family by this pattern. Tt was only in later 



18 



ALAN sugar: the amstrad story 

years, when I was 14 or 15, that I saw the type of industry he 
worked in. They simply told him on Monday morning, "Sorry, 
there's no work. Don't come in." And he used to take it very 
badly.' 

Nathan did his best to cope. He chopped and changed between 
the small garment factories, where conditions were little better 
than in the workshops filled by Jewish immigrants new to London 
in the closing decades of the last century. He went to evening 
classes to better himself, studying the art of cutting a garment. He 
earned a few extra pounds by making children's coats in his spare 
time. Local women would trail their children up to the Sugars' flat 
to be measured for coats for bar-mitzvahs and other special 
occasions. 

The insecurity of his trade fed Nathan's natural caution. Alan 
Sugar, the businessman son whose life has revolved around taking 
risks, looks back on his late father with a mixture of pride at his 
underrated intelligence and exasperation at his failure to grasp 
opportunities. 'He had people rushing after him to make coats for 
them. But the more people he had rushing after him, the more he 
got nervous. He could have opened his own shop. He could have 
made coats. He was a very intelligent man. The problem was he 
would not take a risk.' 

Nathan's responsibilities hung heavily on his shoulders. 'He was 
a provider and he wanted to make sure everyone was provided for. 
He worried very much. He was a very big worrier.' Alan Sugar 
repeats the point, highlighting the facet of Nathan's character 
noted by many visitors to the Sugars' flat. 

Nathan always assumed that if things could go wrong, they 
would do so in spades. He acted as a brake and a dampener on 
his children, flying into a panic if they suggested anything out of 
the ordinary. Nathan struck neighbours like the Kieves as quiet 
and hard-working, if resigned: a man old before his time. But the 
father of the family would show another side in the privacy of 
the Sugars' flat. Nathan had a temper like a slow-burning fuse. 
Almost anything could make him erupt. In one of these moods it 
was not unknown for him to pick up a bottle and throw it against 
a wall. 



19 



DAVID THOMAS 

This moody man formed a close bond with his youngest child 
and later, when he became one of his son's first employees, Alan 
readily put up with his father's little eccentricities. Alan was the 
only one of the children to spend his entire childhood with his 
parents, the others having all been evacuated during the war. As 
the elder children left school and the boom years of the 1950s 
rolled by, the economic pressures on the Sugars lightened a little. 
His parents were able to give Alan slightly more than the others 
had received. 

Nathan Sugar had few means of escape from his tailor's bench, 
but he would regularly turn to the sports page in the paper to 
follow the horses. Even then, however, he was not much of a 
gambler, betting only in shillings. Much of his time was spent 
visiting his family. Unusually for that time, his father, a shoe- 
repairer, had married twice; so too had Nathan's mother. They 
had a family each and a family together, spawning a clutch of half- 
brothers and half-sisters dotted around the East End. 

The Sugars were like many East Enders in basing their social 
life on the family, inviting neighbours into their flat sparingly. In 
this they were typical of most of Northwold's inhabitants, contra- 
dicting the widely held view that East Enders were for ever visiting 
each other. 'These flats are a world of their own. You live a whole 
lifetime here and not know your next-door neighbour,' as a 
character says in Arnold Wesker's play Chicken Soup with Barley, 
which is set in a flat on the Northwold estate. 

The main social occasion in the Sugar household was on a 
Friday night when they would gather to play the card game, solo. 
Nathan was an avid card player and Derek would join in, bringing 
Brenda - his wife-to-be - and friends such as Jeffrey Kieve, who 
lived across the courtyard in the opposite block of flats and was 
later to teach Alan. Alan would be there, hanging around the 
older card-players, but too young to join in the family's activities. 

The Sugars were unusual in one respect in that, unlike many 
Jewish families, they took little interest in their family history. Fay 
and Nathan never discussed whether it was their parents or an 
earlier generation who first came to Britain to escape persecution, 
and their children never thought to ask. The most they knew was 



•20 



ALAN sugar: the amstrad story 

that Nathan's family was originally Polish. No one quite knew 
where their unusual surname came from; one line of thought in 
the family was that it might have been an English version of 
Zuckerman. Alan himself later speculated that his Polish ancestors 
had chosen the name Sugar on arriving in England because they 
worked in a sugar-beet factory. 

Fay's ancestors were thought to be Russian and her maiden 
name, Apple, also to have been Anglicized. The marriage of an 
Apple to a Sugar had raised a smile among the two families at the 
time of the wedding. The Apples and the Sugars lived close 
together in Stepney where Fay's father, Aaron, had his own horse 
and cart to carry goods between the network of small workshops 
in the East End. Fay prides herself on not having worked much 
after her wedding, a pattern common among even the poorest 
Jewish families of those times.. She is blessed with a natural sense 
of fun, a trait shining through a photo which has survived from 
her prime and shows her posing as Mrs Mop to please her 
children: a small, podgy woman with a broad grin clutching a 
mop on the balcony outside her flat. 

The Sugars' lack of intellectual curiosity did not just mark out 
their attitude to their family history. Theirs was not a bookish 
household. They never dreamt of imitating the Kieves' ambition 
of pushing Jeffrey through university. Neither was Nathan 
troubled by the political and trade union fervour which gripped 
many Jews in the garment trade. It is doubtful whether he even 
bothered to vote. 

Their flat could not have been more different from that across 
the way in Weald Square which housed the young Arnold Wesker. 
The same age as Derek Sugar and Jeffrey Kieve, Wesker was 
growing up in a flat full of books, records and friends debating the 
issues of the day. His parents were Communists and working-class 
intellectuals of a type more common in the 1930s. Yet it was the 
Weskers, not the Sugars, who were out of the ordinary, as Wesker 
realized when he went to school. 'Even my Jewish friends saw me 
as way out, bohemian and arty,' he remembers. 

For the Sugars, as for most families on the Northwold estate, 
struggling to make do was enough. The only holidays which the 



21 



DAVID THOMAS 

people of Clapton could afford were occasional day trips to 
Southend and, for some, a working break hop-picking in Kent. 
Northwold was mainly populated by respectable working-class 
families like the Sugars, eager to make the best of what they had. 
Yet there was a fringe element mixed up in the darker side of East 
End life, where fathers were in and out of prison. There were other 
families who buckled under the pressures of living on the bread- 
line, piled up on one another: the thin walls between the flats did 
little to muffle a shouting match or worse between husband and 
wife. 

In this unpromising environment, it was not surprising that 
Alan failed to pass the 1 1 + exam which would have given him 
access to one of the East End's grammar schools. But he did well 
enough to enter Joseph Priestley, a secondary technical school 
catering for children thought to have a technical bent. Pupils there 
were considered a cut above those condemned to secondary 
modern schools. Many of Alan Sugar's schoolmates were destined 
for places at Hackney Technical College and apprenticeships in 
the print, pattern-making and other light industries nestling in 
small workshops and under railway arches all over the East End. 

Set in a crumbling building, Joseph Priestley had an old- 
fashioned head, Jim Harris. A disciplinarian who terrified staff 
and pupils alike, Harris would lecture teachers found playing 
cards in the staff-room for setting a bad example. He modelled his 
methods on the public schools and traditional grammar schools, 
taking much of his inspiration from Merchant Taylors', the public 
school where he sent his own son. He gathered around him 
enthusiastic, well-qualified teachers who tried to set high stan- 
dards for the boys from the estates. 

During Alan Sugar's first two years at the school, Harris was 
preoccupied with a delicate matter. He wanted to ensure that he 
and Joseph Priestley were in the driving seat after the school 
merged with Mount Pleasant, a secondary modern, to form one of 
the new wave of comprehensives. Most of Joseph Priestley's staff 
favoured the then novel concept of comprehensives catering for 
children of all abilities, yet the prospect of having to link up with 
a secondary modern still sent shudders through the staff-room. 



22 



ALAN sugar: the amstrad story 

Jim Harris was determined that they would show the Mount 
Pleasant lot how to run things. 

And so they did. Harris assumed the headship of the merged 
school - named Brookhouse and given a new building a short walk 
down the Upper Clapton Road from the Northwold estate - and 
imposed his well-tested system of prefects and house captains. He 
was so keen for Brookhouse to establish its reputation that he even 
insisted on the staff and prefects wearing gowns. A gesture which 
would be unthinkable in Hackney's schools today, it was calcu- 
lated to appeal to the respect for tradition and ceremony among 
the local people, particularly within the borough's large Jewish 
population. 

Alan Sugar has glowing memories of Brookhouse. The school's 
facilities opened up possibilities which he had never dreamt of: he 
learned to work a lathe, he built a brick wall and he acted in 
Twelfth Night. T could still to this day build a brick wall if I had 
to. And I can still recite parts of Shakespeare. I can turn a lathe 
and read or draw a technical drawing. It was an amazing school. 
It's died off, that kind of school.' 

Yet at this point one of the paradoxes of Sugar's early years 
begins to take shape. No one who came across him as a child 
claims to have noticed anything out of the ordinary in the boy 
from the Northwold estate, not even when they look back with the 
benefit of knowing the shape of his later career. 'If anyone had 
asked who was going to beat the world, I doubt whether anybody 
would have picked Alan Sugar,' says Jeffrey Kieve, who was 
surprised to see Alan when he was ushered in to teach the first 
year in Joseph Priestley as a new teacher fresh from the London 
School of Economics. 

Alan struck Kieve as a determined lad, who conscientiously 
took away his homework in the brown satchel over his shoulders 
and just as conscientiously brought it back the next day. The 
young boy was obviously no fool; he could hold his own. Yet he 
was not one of the adventurous spirits who tested the rules of 
conduct laid down by Jim Harris. Alan chose the science and 
engineering stream, preferring that to commerce as he progressed 
through the school. He enjoyed metalwork, technical drawing and 



23 



DAVID THOMAS 

the labs, yet was not the type of boy who devoted endless hours to 
pulling radios apart. In class, he was neither top nor bottom of the 
form. In short, he was relentlessly average - precisely the kind of 
child overlooked by teachers. 

Vic Pollard, who taught Alan science in the third form and now 
runs an Outward Bound centre, agrees: 'This is one of the strange 
things. Looking back, it just amazes me that this lad has achieved 
so much. He didn't stand out in any way at all.' 

If anything Alan was on the quiet side, a boy who steered clear 
of the school's clubs and other activities. At home, he opted for 
solitary hobbies such as photography and even cookery, making 
cakes and pickling onions. Vic Pollard wondered whether his size 
played its part. Sugar was slight as a boy, a not unimportant trait 
for a lad growing up in a tough area like Hackney. Also he seemed 
a bit of a loner to his teachers, not the type to hang out in a gang. 
T can't remember Alan letting loose that outer shell,' Kieve says. 

Yet the personality of the man who was to become an emblem 
of British business twenty years later was, in fact, taking shape 
under their eyes. Alan began to develop an unusually intense 
interest in business and money. His mother chuckles now over her 
memories of some of the first signs: 'When he was at Brookhouse, 
he wanted to make a school magazine and he wanted a printing 
machine. But he didn't have the money and, do you know what, 
he went to the headmaster and said, "Will you loan me the 
money: 

This early transaction still ranks as one of Sugar's boldest in the 
eyes of someone as unassuming as his mother, not least because it 
worked. Jim Harris, who was later to tell Fay that her son would 
be capable of selling anything from a matchstick to a car, answered 
the budding entrepreneur, 'With your cheek, I will.' 

Alan Sugar's interest in money even predated his days at 
Brookhouse. As a small child, he had made a point of going 
around the flats to ask people if he could take their pop bottles 
back to the shop. The young boy could not understand why adults 
would not make the effort to reclaim the deposits on the bottles 
themselves: 'I'd take twelve bottles back to the sweet-shop down- 
stairs and get a shilling for them. I used to think to myself, "These 



24- 



ALAN sugar: the amstrad story 

people must be mad." All I had to do was walk down the stairs 
with the bottles.' 

Alan later graduated to making pop and selling it to his 
classmates, after a relative had given him a ginger-beer plant. Still 
only 13, he saw it as an opportunity to undercut the price of a 
bottle of Coca Cola or Tizer, which the children from the estate 
could ill afford. He can still recall how to make a gallon of ginger 
beer by feeding the plant every day with two teaspoonsfuls of 
ginger and two of sugar, adding lemon juice and a gallon of water, 
siphoning off the concoction and bottling it up. 

Later in his teens, he would rise at 6 am on Saturdays to boil 
raw beetroot for the local greengrocer. He helped out a family 
friend on his market stall, selling linens and fabrics. Then, for a 
while, he tried selling household goods like washing-up liquids 
from a stall he ran with some other lads on a Saturday. 

Alan even turned a penny from his photography hobby. Old 
friends now in their forties still have black and white snaps of 
family occasions with 'Alan Sugar, Photographer' printed neatly 
on the back and developed in Alan's bedroom-cum-darkroom. 
From a distance of thirty years, he happily admits that his 
photography sideline owed nothing to the technical excellence of 
the shots, but flourished thanks to his instinct for selling. 'People 
want pictures of their kids, always, and grandparents want pic- 
tures of their grandkids more than anything else. So I went 
straight to the grandparents and said to them, "I'll photograph 
your grandchildren for half a crown." And the answer was always, 
"Yes, yes, yes, yes." They could never have enough pictures of 
their grandchildren.' 

Alan showed an unusual talent for selling in his many Saturday 
jobs. Fay Sugar remembers a department store in Hackney being 
so impressed with the schoolboy's prowess in selling shoes that 
they offered him a permanent position on the spot. 

These Saturday jobs were taken on Alan's own initiative; so too 
was his careful husbanding of the money he earned, which he kept 
at home largely unspent. 'I was quite a good saver as a kid. It was 
always indoctrinated into me from my family that you had to save 
your money,' he recalls. 



25 



DAVID THOMAS 

Some of this might be dismissed as the rose-tinted memories of 
proud relatives were it not that Alan also acquired a reputation 
with his teachers. Vic Pollard heard on the school grapevine that 
Alan was involved in all kinds of little deals; he became known in 
school as the boy who could get things for people. This reputation 
was illustrated by the little routine which Alan and Jim Harris, 
the head teacher, would go into whenever they bumped into each 
other. 

'Ah, Sugar, yes, hmm, commerce section, aren't you?' the head 
would say, eyeing Alan from a great height. 

'No, no, sir, science and engineering I'm in,' Alan would reply. 

'No, you're not, Sugar, I know you, you're in commerce,' 
Brookhouse's head would insist. The exchange would end only 
when Alan brought out his school card to prove he was in science 
and engineering. 

The East End has long been famous for its traders. Street 
markets like those in Brick Lane and Ridley Road, the flow of 
goods from the London docks, the myriad small workshops - all 
these contributed to the area's reputation as a place where deals 
were done. Nowadays the head teacher at Homerton House - the 
school which later absorbed Brookhouse - complains that there is 
little contact between his school and business. But in the Hackney 
of the 1960s it was not uncommon for the boys, particularly the 
Jewish lads, to go straight into the family firm. 

Unlike many of his schoolmates, however, there was no relative 
to shape or encourage Alan Sugar's earliest interest in selling and 
trading. Going out to work for someone else was all the family 
understood. 'We were very working-class people,' Daphne 
explains, without the slightest hint of affectation. 

The nearest the family could offer to an entrepreneur was one of 
Fay's brothers, John, who ran a hardware shop in Victoria. The 
shop was known locally for the captions, puns and slogans he used 
to pin to the merchandise; a journalist on a national paper found 
it sufficiently amusing to write an article about his advertising 
techniques. It would make a good story to report that Uncle John 
had held the infant Alan Sugar on his knee while teaching him the 
mysteries of business - a good story, but untrue: Uncle John was 



26 



ALAN sugar: the amstrad story 

no closer than any other relative. Alan Sugar learnt the tricks of 
his trade by himself. 

So Alan already had one step in the world of work as he came 
to the end of his schooling. Ten years earlier, his brother Derek 
had been forced by Nathan to leave school as soon as he was 14. 
Derek had a good head for figures and the backing of his 
headmaster in his wish to stay on at school, with the hope of 
eventually studying book-keeping. But Nathan would have none 
of it, insisting that Derek had to bring in a wage immediately by 
following him into the rag trade as a machinist. An intelligent 
man, Derek quickly grew to loathe the work, drifting after a spell 
of National Service into cab-driving. 

The pattern repeated itself with Alan, who wanted to stay on at 
Brookhouse. The financial pressures on the Sugars had eased by 
the early 1960s, the end of Harold Macmillan's 'never had it so 
good' era, and Nathan felt torn. His gut instinct was for Alan to 
start earning as soon as possible, just like his other children; but 
he also understood that staying on at school might help Alan into 
a better job, maybe even to enter the mysterious world of the 
white-collar professional. 'But apart from my father being very 
nervous, he was also a very sceptical kind of person. He didn't 
have much confidence that anyone from his family would end up 
in anything like university,' explains Sugar. 

An uneasy compromise was reached, Alan staying on beyond 
the minimum school-leaving age to start an A-level science course. 
But in the end Nathan's instinct prevailed. A family member 
recalls a row at one of the family's Friday evening gatherings, and 
that was that. Alan Sugar, aged almost 1 7 and with a few r O-levels 
to his name, abandoned any thought of furthering his education. 
When he left school, he was already earning more from his various 
sidelines than his father. 



The Sugar family's Jewish background reinforced a sense of 
security and solidity in a way which it is difficult for non-Jewish 
people to understand, for they were never meticulous in their 
religious observance. Fay's and Nathan's children were not 



27 



DAVID THOMAS 

brought up to be particularly religious. Only the photos of the 
children's bar-mitzvahs on the walls tell visitors who enter Alan 
Sugar's home today that they are in a Jewish household. 

But it would be easy for outsiders to misinterpret this secular 
attitude as evidence that Alan Sugar has not been deeply touched 
by his Jewish background. Easy and wrong: it has played a part 
in most of the main milestones in his life, outside the realm of 
business. For one thing, all the rites of passage in Sugar's life have 
been carried out according to Jewish custom, from his bar-mitzvah 
in a small synagogue in the Upper Clapton Road to his marriage 
in the Great Portland Street synagogue. Although his friends have 
tended to share his cool attitude to religion, nearly all have been 
Jewish. In his adolescence Sugar made friends with other young 
Jews who gathered around the Jewish youth clubs of Stamford 
Hill, and it was through this network that he first met Ann, his 
future wife. His brother and sisters, and all his friends, also 
married Jewish partners; it would have been 'quite upsetting to do 
otherwise', says his sister Daphne. 

Fay's and Nathan's original move from Stepney to Clapton was 
on the well-trodden route of Jewish emigration out of the tra- 
ditional East End. In the days before widespread car ownership, 
it was dictated by the path of the 653 trolleybus, which ran from 
Stepney to Hackney and Clapton and then on to Stamford Hill. 
By settling along the trolleybus's route, Jewish families could keep 
in touch with relatives and friends in different stages of the exodus 
- a particularly important requirement in the 1930s, when anti- 
semitic agitation by the British Union of Fascists was strong in the 
East End. Alan Sugar moved to Redbridge, on the north-east 
outskirts of London, as a newlywed in the 1960s, the decade when 
Jewish immigration into Redbridge was at its peak. He made his 
first home in Clayhall, the area with the densest concentration of 
Jews. Local Jewish charities were among the main recipients of 
donations from him once he had made his fortune. 

Alan Sugar mixes easily with non-Jews, yet his upbringing in a 
self-consciously Jewish (albeit largely secular) family was an 
important factor in forging his identity. He was named Alan as an 
Anglicized version of Fay's father's name, Aaron; similarly, he 



28 



ALAN sugar: the amstrad story 

named his own eldest son Simon after Nathan's father. Sugar drew 
from his childhood the traditional Jewish emphasis on the family, 
as well as a desire to improve himself. In other circumstances, that 
might have been reflected in his going to university and entering 
one of the professions. Indeed, associates in later life reckoned that 
at times Sugar regretted his inability to pursue his education 
further. Given his milieu, business was the alternative escape 
route. 

Sugar himself is typically unpretentious in trying to explain his 
drive to sell, citing his wish to leave behind the poverty of his 
upbringing: 'My self-reliance really all revolved around finance. 
There was never any question of anything being done for me or 
put on a plate for me. It was quite clear that whatever you did 
and got and achieved, you had to go and do for yourself.' 

Yet many other children shared such an upbringing without 
turning into tycoons. By the time he left school, Sugar had already 
discovered an urge to trade and sell that ran deeper than a simple 
wish for money. He was later to compare this instinct with other 
people's gift for music. T seriously believe I've got an inborn 
talent,' he explained. Tve got an inbuilt aptitude for trading and 
dealing and scenting the way the wind is blowing.' 

However, the young Sugar had to put his entrepreneurial 
instinct on ice as he emerged into the adult world. 



It is difficult to imagine anyone more unsuited to the daily life of 
the Civil Service than Alan Sugar. The cautiousness, the anonym- 
ity, the respect for established procedures and patterns of thought, 
the routine meetings - everything that marks out the civil servant 
is anathema to him. Yet it was to the statistics department of the 
Ministry of Education and Science, one of the most bureaucratic 
of all the Whitehall departments, that he headed for his first job. 

Unsurprisingly, Nathan Sugar encouraged his son in this move, 
seeing an office job in Whitehall as a step up the ladder into a 
professional career. For Nathan, keeping his youngest child at 
school beyond the school-leaving age had been both a sacrifice 
and an investment. 'If I had stayed at school longer than I needed 



29- 



DAVID THOMAS 



and then ended up working as a salesman in a shoe-shop or a 
men's-wear shop, it would have seemed a total waste to him,' 
Sugar explains. 

Knowing no better himself, Sugar went along with his father's 
wishes. In any case, working on statistics in the Ministry respon- 
sible for science appealed to the young Alan. He had been so 
pleased to get his first job that he had not enquired too closely into 
the nature of the work. 'Science - this was something I had always 
been interested in. Statistics, maths ~ I wasn't too bad at that. So 
I thought I'd go for it.' 

Sugar was in for a shock. Far from being scalded by the white 
heat of the technological revolution, which Harold Wilson had just 
sold to an expectant nation during the 1964 general election 
campaign, he found himself shuffling paper. The most intellec- 
tually challenging work was coding research for the Plowden 
Inquiry into primary education: how many children drank milk in 
the morning - and how many did not? 'When I got there, I found 
it was the most unbelievable bore going,' Sugar says. 

People dealing with him in later years were invariably struck by 
two interrelated facets of his intelligence: the speed with which he 
digests complicated proposals and his tendency to become bored 
very quickly. Already an impatient man, he found the Ministry a 
nightmare and still sounds angry when he remembers it. 'You 
literally had to watch the clock. It was total agony waiting for that 
clock to tick to five o'clock.' 

During his brief sojourn in Whitehall, much of Sugar's energy 
went into his after-hours activities. He was probably alone among 
his new colleagues at the Ministry in holding down a Saturday 
job, first in a men's-wear shop in Tottenham Court Road, then in 
a chemist's near Walthamstow market. 

Sidelines like these Saturday jobs brought in extra cash and 
helped to satisfy Sugar's drive to sell, but they were not enough to 
compensate him for the boredom and lack of prospects at the 
Ministry. 'He didn't like it because it was a sitting-down job,' is 
his mother's wry comment on this interlude. 

It took Sugar less than a year to realize that he would always be 
a square peg in the world of committees. Yet when he decided to 



30 



ALAN sugar: the amstrad story 

leave, he was persuaded by his father's anxious vision of a 
respectable white-collar future to try something similar. He 
entered the statistics department of Richard Thomas & Baldwin, 
a steel company central to Britain's heavy industry-dominated 
economy of the time, but long since a victim of rationalization. 

The tedium of the work, which he soon found as dull as anything 
at the Ministry, was relieved by the office crowd who took a liking 
to the young East Ender. They joked a lot, and shared with Sugar 
their worldly wisdom. Already too tied down by wives and 
mortgages to think of quitting themselves, they reckoned he still 
had a chance and were sure he was in the wrong job. They told 
him that selling was where money was to be made: 'Get out there 
and get into sales.' 

Telling Sugar to sell was like telling Rudolf Nureyev to dance. 
In the lunch-breaks from his Saturday job in the chemist's shop in 
Walthamstow, Sugar used to meet a friend called Malcolm Cross 
who worked locally as a television engineer. He had got to know 
Cross - like himself from the poorest of East End families - a 
couple of years earlier in one of the Jewish youth clubs in Stamford 
Hill. They were part of a shifting gang of Jewish teenagers from 
Hackney, who made occasional forays to the coffee bars and ten- 
pin-bowling alleys of the West End. 

Between them, Cross and Sugar hatched a scheme which can 
lay claim to being Alan Sugar's first commercial enterprise as an 
adult. The idea was to combine Cross's skills as a television 
engineer with Sugar's already evident talents for selling. The pair 
of them would buy run-down, second-hand TV sets in two-dozen 
batches; Cross would do them up; and Sugar would sell them. 

By the mid-1960s, TV had become a necessity for even the 
poorest Hackney home, yet the sets on offer then were unreliable, 
forever breaking down. The Hackney Gazette had a regular slot for 
adverts from friendly, local TV repair men, and also advertised 
reconditioned TV sets from as little as £5. 

Sugar added a little twist to this well-established part of 
Hackney life, advertising Cross's reconditioned TVs as presents 
given to the Sugar family which they did not happen to need. 'He 
used to sell them in his bedroom, pretending they were a one-off, 



31 



DAVID THOMAS 

unwanted gift. As he sold one, we used to put another in the 
bedroom,' explains Cross. 

Pretty soon, the whole family was aware of Sugar's new sideline. 
Derek, Sugar's brother, stopped his cab one day at a coffee stall in 
Stoke Newington. 'I saw a little advert pinned on it, "TV for sale", 
and I did a double take because it was my mother's phone 
number.' 

Daphne became used to the procession of odd people climbing 
the stairs of Woolmer House to look at the 'old monstrosities', as 
she referred to the TVs. At times Fay had to field the complaints 
of the not inconsiderable flow of dissatisfied customers: 'They used 
to come and say, "Television's not working," so I'd say, "Well, 
here's the money back.'" 

Sugar and Cross took their fledgling business seriously, christen- 
ing it 'Maurann', an amalgam of the names of Cross's wife 
Maureen and Sugar's girl-friend, Ann. Sugar arranged headed 
notepaper for Maurann and the partners hired a room where they 
stashed their TVs. The enterprise lasted a year at most, but it 
represented a bridge between Sugar's brief days as a salaried 
employee and his life as an entrepreneur. 

Selling fitted Sugar's natural temperament better than life in an 
office ever could, but his eventual decision to quit Richard Thomas 
& Baldwin was characteristically dictated by something much 
more immediate: 'What I was really after was wheels.' 

He had had plenty of time to contemplate the freedom offered 
by a car as he looked out from the Northwold estate on to the busy 
Upper Clapton Road, scything its way through Hackney. A car 
had always been way off the map for the Sugar family. 'A car was 
considered to be an absolute luxury. Rich people had cars - that's 
how you viewed it,' Sugar says. 

The earliest ambition that Sugar can remember was for his own 
car. And since neither he nor his family could afford one, the only 
way of realizing this ambition was to find a job which came with 
'wheels'. 

Sugar answered an advert placed by Robuk Electrical, then a 
well-known tape recorder dealer based in the Holloway Road in 
North London. Robuk was adding to its firm's sales force throughout 



32 



ALAN sugar: the amstrad story 

the country and Sugar was given the job of covering London, selling 
its tape recorders to retailers all over the capital. Nathan was less 
than impressed by his son's move. 'My old man was going mad. I'd 
had three jobs in about nine months, while he tried to stick with one 
job for 15 years,' recalls Sugar. 

In fact, the job was perfect for him. It allowed him to roam 
London by himself, getting to know the myriad outlets for electri- 
cal goods; it gave full vent to his selling instincts; and, crucially, it 
gave him the use of a van. 

Throwing himself into the work, Sugar was delighted to hear 
one day that Robuk had done a deal with Currys, the electrical 
goods chain, which would allow him to sell tape recorders direct 
to each Currys store. Currys's shop managers had previously 
lacked the authority to order tape recorders without contacting 
head office. Suddenly they were set free, an event which Sugar still 
remembers as a tremendous coup: 'Well, to me that was like a 
licence to print money.' 

Sugar talked every Currys store in London into buying Robuk's 
tape recorders, but come pay day, he found that his commission 
was about a quarter of what he had been expecting. None too 
pleased, he demanded an explanation from Robuk and was told 
that because the Currys business represented a bulk deal, the 
commission was correspondingly less. There was a row and, right 
or wrong, Sugar waved goodbye to Robuk. 

His next port of call was R. Henson Ltd, an electrical wholesaler 
in Finchley. Hensons handled the products which were pouring 
into London from importers and manufacturers to fuel the mid- 
1960s boom in electrical goods: transistor radios, car aerials, car 
radios, clock radios, walkie-talkies. Sugar's job was to take samples 
to Hensons's retail customers, strike a deal, then deliver the order 
and collect the payment. 

Driving around in the Hensons van, Sugar rapidly boosted his 
knowledge of the hundreds of independent outlets and shifting 
product ranges that made up the electricals business. But respect 
for his new employer did not develop in equal measure. 

'Hensons was really an eye-opener. I thought they were a stingy 
load of sods, I did. I used to sell the stuff for them, then deliver 



33 



DAVID THOMAS 



the bloody stuff, then collect the money for them. And you never 
got a pat on the back for doing anything right,' he says. 

The final straw came when Sugar located a stock of surplus 
records, a line which had sold well for Hensons in the past. He 
told them he could buy these cheaply and already knew someone 
who would take the records off their hands at a tidy profit. 
Hensons gave him the go-ahead and the cash to cover the deal. 
Sugar duly returned with a couple of hundred pounds to show 
from a day's work. 

'If Hensons had been clever at that stage, they would have 
made me a partner of the firm, and it might have been Hensons- 
cum-Amstrad by now,' Sugar says. 

Instead, they criticized him for not making an even bigger profit 
and, adding injury to insult, refused to give him an extra bonus on 
the deal. That naffed me a little bit,' Sugar adds with feeling even 
twenty years later. Peter Henson, who now runs the family firm, 
does not remember this incident, but Sugar continued to trade 
with Hensons after he set up his own business, suggesting that 
relations were soon restored. 

This episode crystallized the suspicion which had been forming 
in Sugar's mind that he could do better on his own. He used to 
pick up goods for Hensons from one of its main suppliers, 
Binatone, a company which had been importing anything which 
would sell from the Far East since the late 1950s. With the other 
drivers, he would queue at the loading bay at Binatone's main 
premises in Finsbury Square, just north of the City. 

During his enforced waits by the loading bay, Sugar fell to 
chatting with Gulu Lalvani, one of the two brothers who runs 
Binatone. Although only a few years older than Sugar, Lalvani 
was already a suave and successful businessman, an obvious role 
model for the ambitious East Ender. A tall, elegant man who set 
up Binatone in 1958 after coming to Britain from Bombay, Lalvani 
remains one of Sugar's closest friends. He remembers Sugar's 
acumen and will to succeed marking him out from the other 
drivers. 

It had not taken Sugar long to understand the fundamentals of 
the business. He knew the credit terms Binatone gave Hensons; he 



34 



ALAN sugar: the amstrad story 

knew what Hensons paid for Binatone's goods; he knew Hensons's 
outlets and the prices they paid. In short, he grasped that Hensons 
was getting goods on credit and selling them for cash: 'In the end, 
I realized that's all they were - a glorified buying and selling 
agency.' 

Sugar was earning about £20 a week with Hensons, and he 
reckoned he could make that in a day on his own. He spent £80, 
the bare minimum, on a second hand Mini-van and went to 
Lalvani with a proposition. He was going to quit Hensons, so if he 
gave Lalvani a seven days' post-dated cheque for a couple of 
hundred pounds, would Lalvani give him goods to that value? 

Lalvani hesitated. In those days John Henson was one of his 
best friends and the Binatone chairman was loath to be seen to be 
encouraging one of Hensons's workers to strike out in competition 
with him. 'Look, I'm going to quit, whether you help me or not,' 
Sugar said. 

To satisfy his conscience, Lalvani suggested as a compromise 
that Sugar should go away for a couple of weeks' holiday before 
launching out on his own, so that it did not look as though Lalvani 
had directly encouraged him to make the break. Sugar agreed. 

When the two weeks were up, Sugar duly presented Binatone 
with the post-dated cheque and collected his goods. Lalvani was 
surprised to see Sugar's van rolling into Binatone's premises that 
same evening: 'Give me the cheque back. Here's the cash,' Sugar 
said, handing over the money. 

Sugar had managed to sell the week's deliveries in a day. He 
was back at Binatone bright and early the next day for another 
consignment. Sugar earned a profit of 20-25 per cent on every- 
thing he sold. It was 1966 and Sugar was a 19-year-old entrepren- 
eur. He knew at once that he had made the right move. 

That was not how it appeared to Nathan, the ever-cautious 
father who had spent his whole life working for other people. 
When Sugar told Nathan that he was quitting Hensons to work 
for himself, his father asked anxiously, 'But who's going to pay 
your wages?' 



35 



Starting Out 



The year 1968 has gone down in the history of the twentieth 
century as a year of pseudo-revolution. Students revolted in Paris, 
occupied university buildings in London and held love-ins in San 
Francisco. For a brief moment, the foundations of Western society 
seemed to be cracking as the children of the privileged attacked 
authority in the name of the working classes. 

The year also saw momentous changes for Alan Sugar, but for 
the young East Ender the upheaval was a private matter. A 
product of the working class himself, Sugar had no time for the 
romantic myths of his middle-class contemporaries. He knew that 
changes in his life would flow from his own efforts. And in 1968 
the changes to Alan Sugar came thick and fast: he married, he left 
Hackney and he founded Amstrad. 

Sugar had met Ann Simons through the network of friends 
which congregated around the Stamford Hill youth clubs. A shy 
girl a couple of years younger than Alan, Ann soon formed a 
strong attachment to her rough diamond boy-friend. Sometimes 
when she had a day off from her job as a hairdresser, she would 
accompany him on his deliveries, quietly reading a book in the 
front of the van while he loaded or unloaded a consignment of 
aerials or radios. 

The budding romance was not greeted with universal approval. 
At first the Simons family did not think much of Ann's choice. 
What could Sugar offer her? He had already gone through a string 



36 



ALAN sugar: the amstrad story 

of jobs and was now running a one-man business from the back of 
a clapped-out van. He was not much of a catch for a daughter 
who might have hoped to marry a lawyer, an accountant or a 
doctor - someone with a future. 

Not only did Sugar have nothing to his name, he even came 
from the wrong part of town. Living in Redbridge - the London 
borough to the north-east of Hackney where many Jews moved 
once they were prosperous enough to leave the East End - the 
Simonses had no wish to be linked with a family still stuck in 
Hackney. Johnny Simons tried to impress his views on his 
daughter, but to no avail. 

Sugar was already acutely aware that he came from the wrong 
side of the tracks. As a teenager, he had begun to make friends 
outside his immediate social circle, partly through the Jewish 
youth clubs in Stamford Hill - an area one step further away from 
the East End than Clapton and visibly more affluent. He socialized 
with people of his own age from Stamford Hill or even, like Ann, 
from Redbridge. 'I met a different class of person. Some of their 
parents were in business and I saw a little bit of a different way of 
life,' Sugar recalls. 

He was not the only member of the gang to come from the poor 
parts of Hackney; Malcolm Cross, his partner in the television 
venture, shared this background. But Maureen, Cross's future 
wife and also part of the crowd, distinctly remembers: 'Alan was 
the poorest out of all the boys. He really was. Few of us had 
anything, but he had less than all of us.' 

Some of his new friends lived in houses which seemed way 
beyond the dreams of the Sugar family. This disparity fed his 
drive to leave behind the poverty of Hackney, which had been 
growing more intense in the young Sugar: 

As I began to mix with this different circle of friends that came 
from better-off families, it was evident there was no point asking 
my father if he could buy me some nicer clothes, a car and 
things like that. It was not that he didn't want to, he couldn't. 
So it was clear that if I ever wanted anything, I had to do it for 
myself, simply to improve my lifestyle. So that's what I did. 



37 



DAVID THOMAS 

The idea of dropping Ann because he was too poor cut right across 
the grain of Sugar's sensitivities. Disapproval from his girl-friend's 
family only reinforced his determination to make a go of his one- 
man business. Associates of the time, like Binatone's Gulu Lalvani, 
noticed the impact on Sugar: Tt motivated him even more. He 
wanted to prove to his girl-friend's father that, O.K., he might not 
be a doctor or an accountant or a lawyer, but he could be a 
successful businessman,' Lalvani recalls. 

In the end, the family opposition proved futile and the young 
couple were married in 1968, when Sugar was 21. The families put 
on a good show at the wedding, a clear signal that the Simons side 
was reconciled to Ann's choice. The service in the Great Portland 
Street synagogue was followed by a reception in the West End. 

Alan and Ann Sugar bought a house in Marlands Road, 
Clayhall, an area of Redbridge not far from where Ann's parents 
lived, reflecting the fact that Sugar's business was already showing 
signs of success. Clayhall is full of neat semi-detached houses 
newly clad in stone chippings. The classier of the dwellings sport 
mock-Tudor exteriors, with black and white panelling and opaque 
diamond window-panes - or even Greek frontages, with fake Doric 
columns supporting the porchways. 

The young couple could not yet manage anything quite like 
that, but they moved into a quiet street of handsome semis with 
wide bay windows. In any case, it was enough that they were in 
Clayhall at all; they were on the first rung of the ladder. 



It was not all plain sailing for the young businessman, however. 
Sugar had had to struggle with his Mini-van from his earliest days 
working for himself. It had been the cheapest which he could find 
and as a result was for ever breaking down: 'Because it was always 
in the bloody workshop, I couldn't go out and sell. We needed a 
more reliable vehicle because that was the lifeblood,' Sugar 
explains. 

Under normal circumstances, the solution would have been 
simple: Sugar would have bought a new van, using hire-purchase. 
But when he first went into business, he was too young to sign a 



•38 



ALAN sugar: the amstrad story 

hire-purchase agreement. So he turned to his parents for help, but 
in doing so once more ran up against his father's ultra-cautious 
instincts. 

Sugar's idea was that he would put down a £100 deposit - the 
most he could manage - towards a Vauxhall Viva costing about 
£500. Nathan would sign the hire-purchase agreement and Sugar 
would then take care of the weekly payments. But Nathan felt 
anxious about the prospect of his son being in hock to a hire- 
purchase outfit, so instead he came up with the balance of the 
money from his own savings and Alan paid off a weekly amount 
to his father. Sugar recalls: 'It was the typical mentality of my 
father. He couldn't understand that I'd ever be able to pay the 
money back.' 

Robbery was the next hurdle Sugar had to confront. At the 
time, he was keeping his stock in a lean-to built for the purpose on 
the back of the Sugars' new house in Marlands Road. Describing 
the setback to his fledgling business, he says: 'I'd built the business 
up and was making a good salary and had about £1,500 worth of 
stock. And one day someone broke into the back of the house and 
nicked it all. It was bloody stupid when you think about it.' 

Luckily for Sugar, the robbery was not fatal to his business 
because a lot of his assets were in the form of money owed to him 
by customers. But he quickly learned his lesson: 'I wasn't com- 
pletely flattened, but I decided then that I needed some premises, 
a burglar alarm and proper insurance, so that it could never occur 
again.' 

Sugar found his first business premises in a house in St John 
Street, just down from Islington High Street in North London and 
almost opposite the Sadler's Wells opera house. St John Street 
marks the point where Islington's bistros and antique shops give 
way to an area of dingy small businesses and workshops. His 
headquarters was in a tall, austere terrace of houses which would 
have made fine homes if they had been tucked into an Islington 
back street. But since the terrace was on the main road to the City 
and Fleet Street, it was in fact made over almost exclusively to 
small businesses. Yet the aesthetics of the building were Sugar's 
least concern; what mattered to him was that the larger premises 



39 



DAVID THOMAS 



allowed him to build up bigger stocks and a wider range of 
products. 

Through a cousin of Ann's, Sugar found his first accountant - 
Guy Gordon, a podgy, avuncular man who was a partner in a firm 
specializing in small- and medium-sized companies. Gordon told 
Sugar what records to keep and, just as crucially, initiated the 
young businessman into the mysteries of limited liability. Shortly 
after Sugar moved into Marlands Road, Gordon advised him to 
form a proper company so as to guard himself against bad debt 
problems, a perennial risk in his line of territory. By doing so, 
Sugar would limit his personal liability if his company ran into 
difficulties and so prevent his house - his one personal asset of any 
value - becoming vulnerable to creditors. 

And so in 1968, on the first day of November, A. M.S. Trading 
Company (General Importers) was registered as a limited com- 
pany. The name is instructive for two reasons. First, Sugar's 
initials are spelt out in the title, showing it to be a half-way stage 
on the road to 'Amstrad', the brand which was to become a 
household name almost two decades later. 

Second, the use of the term 'trading' - and even more so of the 
description 'general importers' - reveals how Sugar envisaged his 
business at this stage. It was to be a buying and selling operation: 
manufacturing was not part of the equation. And since he was 
mainly buying and selling electrical goods, much of his merchan- 
dise would inevitably be imported. 

Sugar slipped into using 'Amstrad' not long after the registration 
of his company, certainly well before he formally changed the 
name of the operation to A. M.S. Trading (Amstrad) in November 
1972. As he began to take a regular flow of goods from particular 
importers, it occurred to him that he would create a more 
substantial image among his retail customers if he had his own 
brand name. Tt just seemed that if I had my own brand name on 
a product, it gave it more credibility, rather than it being simply 
something passed on by a wholesaler,' he explains. 

Cigarette lighters and intercoms for connecting different rooms 
in an office were the first products to bear the Amstrad name. 
Sugar did not manufacture these: they were goods which he 



40 



ALAN sugar: the amstrad story 

bought from importers and then badged. Amstrad - a name which 
is so catchy that it deserves to have been invented by a team of 
image consultants - simply evolved as a shortened version of 
A. M.S. Trading: 'It sounded right, but it was more luck than 
judgment. A lot of the other brand names of people we were 
dealing with in those days were also anagrams of their trading 
titles,' Sugar recalls. 

Sugar was just 21 when he registered his company. That same 
year, a man seven years his senior and destined to play a 
considerable role in the Amstrad story was making his first big 
splash. Clive Sinclair caused a considerable stir by placing a four- 
page colour ad for an expensive amplifier in Hi-Fi News, the 
normally black-and-white journal of the audio hobbyists. Sinclair 
was becoming an exciting name in the world which Sugar would 
soon enter himself. 

The next couple of years were punishing ones for Sugar. He 
spent the time remorselessly extending his contacts with the 
multitude of small independent outlets which dominated the 
electrical goods and audio businesses at the time. That meant 
buying from whichever wholesalers were prepared to shave a few 
pence off their cost, and selling to whichever outlet offered the best 
price. He had to make countless journeys in his van, not just all 
over London but also to towns like Cambridge and Portsmouth - 
wherever, in fact, the best deal could be done. 

One of the first people he came across after moving into St John 
Street was Colin Lewin, who had a shop trading in electrical goods 
in Gray's Inn Road, just down from Sugar's premises. A few years 
older than Sugar, Lewin had the same short, stocky, bustling 
physique as the younger man and shared something of his 
background too. Brought up in a poor family, although on the 
other side of London in Battersea, Lewin had originally worked as 
a telephone engineer for the Post Office in the West End, but this 
left him with so much time on his hands that he started selling 
transistor radios while waiting to be called out to repair a 
customer's telephone. 

Lewin and Sugar were in the same line of business, selling 



41 



DAVID THOMAS 

amplifiers, tuners, turntables, car radios, aerials, speakers, transis- 
tor radios ... in fact, pretty much anything they could get their 
hands on. Both young and ambitious, they soon started trading 
with each other - swapping bits and pieces they had picked up, 
sometimes in batches of a hundred, sometimes in one-offs. 

The two became friends and Sugar often stopped off at the cafe 
next to Lewin's shop for a cup of tea - provided, that is, that he 
could carve out the time, for his first priority was always doing a 
deal. 'Alan certainly put in a good six-day week. He'd sell 
anything, anywhere,' Lewin remembers. 

Sugar took more and more of the radios, tape recorders and car 
aerials imported from the Far East by Gulu Lalvani for Binatone, 
impressing him with his ability to spot the slightest opportunity. 
One day, for instance, Lalvani happened to mention that Binatone 
had imported a batch of radios with a slight fault; although the 
repair would be easy to do, his engineers were preoccupied with 
other work, so the radios would have to lie around in the 
warehouse for a few months. 

'Give me batches of 100 to 200 a time and I'll get them done for 
you,' was Sugar's immediate reaction to the news. So he started 
taking batches of these radios home overnight and bringing them 
back adjusted the next day. It was pretty clear to Lalvani that 
Sugar and his wife were doing the work themselves; this was the 
kind of initiative that he was beginning to expect. 

At the end of the 1960s Tottenham Court Road, in London's 
West End, boasted the greatest concentration of independent 
audio and hi-fi shops in the capital, just as it does now. Twenty 
years ago, there were few more important people in this business 
microcosm than Ronnie Marks, boss of Premier Radio which had 
started as a wholesaler and then became one of the first outfits to 
open an audio shop on Tottenham Court Road. Marks was 
working in the back of his wholesale operation one day when the 
chap on the counter came in and said, T've got a new face out 
front. He says he needs some credit.' 

Marks went out to see this new face and found a brash youngster 
who said his name was Alan Sugar, he had just started up on his 
own and he needed some credit to get going. Marks was cautious: 



42- 



ALAN sugar; the amstrad story 

'Credit is difficult. We don't know you or what you're like.' Sugar 
was adamant: 'If I don't have credit, I can't get business.' 

After a bit more sparring, Marks agreed a 'one-over-one' 
arrangement, the most basic type of credit. Sugar could have a 
load of goods from Premier, but he had to pay for them when he 
came back for another consignment. It was take-and-pay and 
Sugar started coming into Premier every Friday, increasing his 
take week by week. After a few months testing him out on this 
tight leash, Marks agreed to loosen the terms and gave Sugar a 
month to pay for each batch of Premier's goods. 'He always, 
always honoured his obligation,' Marks recalls. 

Credit from established wholesalers like Binatone and Premier 
Radio helped Sugar to generate credibility with other suppliers, 
who would ring up Lalvani or Marks to check out his record 
before giving him credit themselves. Sugar was learning how to 
balance his sources of credit and cash, allowing him to finance his 
growing business by building up a range of credit sources. By 
selling quickly, he was taking in money on goods he had not yet 
paid for himself. 

Still only just into his twenties, Sugar was rapidly forging a 
reputation as one of the most promising young middle men in the 
business - a reliable conduit of goods from the wholesalers to the 
independent retailers. Unknown to either group, however, he was 
also hedging his bets by taking a stake in a retail operation himself. 



Home in Clayhall, Alan Sugar had struck up a friendship with a 
man of his own age called Ashley Morris. The Sugars were linked 
to Morris through one of those family chains typical of the 
Redbridge Jewish community: the first cousin of Morris's then 
wife was married to Sugar's sister Daphne. 

The friendship blossomed in 1969 when the birth of the Sugars' 
first son, Simon, coincided with that of Morris's daughter. The 
two wives, having gone through ante-natal clinic together, started 
dropping in on each other during the day and it was not long 
before the couples were going out as a foursome in the evenings. 

The two men took to discussing business over dinner and Sugar 



43- 



DAVID THOMAS 

explained that one of his old friends, Malcolm Cross, had opened 
a discount hi-fi shop in Stamford Hill, called Audio Supplies. 
Customers would travel from all over London in the hunt for 
bargains on the latest amplifier, tuner or other hi-fi item. Cross 
and his partner were also running a mail order operation by 
advertising lists of products in the hi-fi magazines. Sugar told 
Morris that he would be amazed at the amount of money which 
was changing hands in Cross's shop. 

Morris worked in the fashion business in those days, but Sugar 
persuaded him to cast an eye over Audio Supplies one Saturday 
morning. He introduced Morris as his cousin because he did not 
want Cross to realize they were there to take a look at the business. 
Sugar and Morris made some small talk and saw that, sure 
enough, people were coming in every five minutes and buying 
£100, £150, £200 worth of equipment — money to be reckoned with 
at the end of the 1960s. 

Back home, Morris agreed to Sugar's suggestion that they 
should open a similar operation themselves. They took premises 
in Pentonville Road, not far from St John Street, called their 
business Global Audio and drew up adverts for the hi-fi magazines 
modelled on Malcolm Cross's efforts. Morris explains: 'We 
decided to set up a business which was really copying Audio 
Supplies completely, except we slashed the prices as much as we 
could, while still making a margin to live on.' 

Sugar and Morris put in £2,000 each to launch Global Audio, 
but the business got off to a bad start when a company from which 
Sugar had ordered £1,700 of record decks went into liquidation 
just before making the delivery. That wiped out almost half of 
Global Audio's initial capital. Sugar retrieved the situation by 
arranging a line of credit from one of his many contacts in the 
business, Dallas Electrical, an outfit based in the City which sold 
hi-fi gear and musical instruments. 

The two partners were listed as directors of Global Audio, but 
Sugar wanted his involvement in the shop kept quiet. He feared 
that other retailers would resent it if they realized that Sugar - 
whose prime job was to sell to them - also had a retailing 
operation: they would suspect that he was selling equipment to 



44 



ALAN sugar: the amstrad story 

Global at discount prices. Manufacturers might also take a dim 
view of Sugar's move; in those days they tended to deal only with 
wholesalers, not with retailers, so they might accuse him of 
confusing the market by dabbling in both sides of the business. He 
was just becoming established in the close-knit London audio 
world and could not afford to put his hard-won position at risk by 
being seen to ride two horses at once. 

Sugar would drop into Global Audio on a Saturday, chat to the 
kids in the stock-room, serve a customer or two and pick up his 
cut of the week's takings. His position in the company was never 
broadcast, although plenty in the trade knew about it. Colin 
Lewin, for one, would wander up from his shop in Gray's Inn 
Road to Global Audio on a Saturday and find Sugar serving 
behind the counter: 'Alan was like me. He couldn't just sit there 
and watch people being served.' 

But Sugar did not like having to keep his stake in Global Audio 
perpetually hidden. He told Morris: 'Look, I can't stay involved 
in this much longer. I want to go places in the hi-fi world. So you 
can buy me out.' He put a take-it-or-leave-it offer on the table, 
which Morris took. 

By then, Sugar had another reason for wanting out of Global: 
he had launched the first Amstrad product and it was beginning 
to take off. 



By 1970, when his company was almost two years old and Sugar 
was 23, he was already becoming aware of the limitations of his 
chosen field of operations: 

I rapidly started realizing that I was just going to be a dealer 
for the rest of my life. It wasn't really exciting because there was 
no big money to be made. Everything I was selling was 
commodity stuff. There were no big coups to be had, because 
everybody knew the prices. 

Casting around for something to drag him out of the one-man- 
and-a-van rut, Sugar concluded that he could move his business 



45 



DAVID THOMAS 

on to a new level if he could find a product which he could 
manufacture more cheaply than anyone else. The problem was: 
what? 

In the early days of hi-fi, it was standard practice to buy all the 
parts that made up a hi-fi system separately. Customers would 
often purchase a record player in one place, amplifiers elsewhere, 
turn to yet a third brand for speakers and then assemble the 
components into one system. Hi-fi enthusiasts would also have to 
acquire some distinctly unglamorous, low tech parts before they 
could unveil their system to an admiring world: one was the 
wooden base on which a record player sat, known in the trade as 
a plinth; the other was a plastic cover to keep dust off the turntable. 

Sugar noticed that the cost of the plastic dust covers being sold 
at the time appeared inordinately high. He decided to find out 
why and soon learned that the covers were made by a process 
known as vacuum forming. A sheet of acrylic - or perspex, as it 
would be called today - was sucked into a warm mould, shaped 
and then trimmed off. The trouble with this process was that it 
used a lot of acrylic, a relatively high-cost raw material. 

The young businessman reckoned that there must be a better 
way of doing it, so he picked up the phone book and started 
ringing any organization he could find with the word 'plastic' in 
its title. He discovered that it would be possible to make the covers 
differently, using a technique called injection moulding. This 
would require a large initial investment in an injection moulding 
tool, but would yield big savings on the material costs. 

After another burst of phone calls, Sugar found a tool-maker 
who could produce the injection moulding tool for him. He 
invested £1,805 in the tool, which was a fortune to him at the time, 
but that proved the easy part: T took the cheapest bloke going. 
There was heartache and aggravation there, because he was 
supposed to make the tool in three months and it took about five 
months. And when he made the first moulds, they broke. He 
moulded them at the wrong temperatures,' 

But Sugar persevered and soon he was making plastic dust 
covers more cheaply than anyone else in the business, buying in a 
set of plinths and sticking them together. Plinths and covers were 



46 



ALAN sugar: the amstrad story 

the first articles which he manufactured himself. They were 
products which everybody wanted. He shifted his plinths and 
covers in large quantities, cashing in on the contacts he had made 
and the reputation he had built for always honouring deals during 
his van delivery days. 

Sugar sold his plinths and covers to the cluster of Tottenham 
Court Road hi-fi specialists like Premier Radio and Lasky's. He 
sold them to manufacturers such as BSR, who incorporated them 
into their ranges. He sold them to Ashley Morris at Global Audio, 
who advertised them at £2.95 each. He sold them at a slight 
discount to other old pals, like Colin Lewin who at one stage was 
taking 1,000 a time from Sugar and passing them on to his dealers 
in bundles of a hundred. With everybody screaming for his covers, 
the injection machine at the moulder's plant could not pump them 
out fast enough to keep up with demand. 

With the benefit of a large amount of hindsight, Sugar's plastic 
dust covers already exhibited many of the features which were to 
become Amstrad hallmarks. The product might not itself be 
innovative, but Sugar had broken new ground by radically reduc- 
ing the cost of its production. This allowed him to carry off his 
three-card trick: high volumes, low prices and large margins. He 
could offer his customers so keen a price that it was bound to 
undercut his competitors, and he could simultaneously generate 
high profit margins for himself. 

Of course, this was not how it appeared to Sugar back in 1970, 
but he was fully aware that his business was set on a new 
trajectory: 'We came out of the era where I was just distributing 
all those wholesale imports into an era of becoming a 
manufacturer.' 



One evening, the Sugars and the Morrises went to a smart West 
End restaurant, where over dinner the two men fell to chatting 
about their favourite subject. Ashley Morris opened up: 'I don't 
care what you say, Alan, success in business is 50 per cent hard 
work and 50 per cent luck. You have to have the luck running 



47 



DAVID THOMAS 

with you, otherwise you get nowhere.' Sugar's reply was immedi- 
ate: 'No, mate, it's hard work. It's all hard work.' 

In his early twenties, Sugar had reason to be pleased with the 
progress of his business. He had disproved all the sceptical 
forebodings of his father by showing that he could live off his wits; 
he had left Hackney behind, moving into an area inhabited by the 
most affluent of his teenage friends; he was earning enough to keep 
him and his young wife - in fact, more than enough, since they 
were beginning to be able to afford the occasional extravagance, 
like a night out at an expensive restaurant or discotheque. 

Yet Sugar displayed no sign of throttling back or of taking it 
easy. On the contrary, friends of his from that time describe an 
intense man, preoccupied with succeeding in his business to the 
exclusion of everything else. He confided in one friend that the 
demands of running his business on his own used to keep him 
awake at night. He had a compulsion to win which was quite out 
of the ordinary, as one old friend relates: 'If you sat down and 
played a game with him, if he wasn't winning, he played till two 
or three o'clock in the morning till he was winning. He was like a 
bear with a sore head if he didn't win. He wanted to prove 
something all the time to everybody.' 

True, Sugar liked few things better than going out with his 
friends for a meal in the evening, an event which would have 
seemed an unimaginable luxury for him only a few years earlier. 
But even on these occasions he found it difficult to throw off the 
cares of his work and became really animated with his friends only 
when he was discussing business. T can't remember him relaxing, 
to tell you the truth. I can't visualize it,' says one old friend. 
Another adds: 'He never seemed as if he was listening to you. His 
mind was always ticking over on something else.' 

The fact is that with his growing business to manage, and the 
responsibility for a young family, at this stage of his life Alan 
Sugar had no time left over for demanding hobbies or interests. 
Daphne, his sister, remembers from this time: 'Once he started in 
business, that was his hobby.' 

Daphne was close to the Sugars while they lived in Marlands 
Road. Although she did not realize how well Alan Sugar was 



48 



ALAN sugar: the amstrad story 

doing in his business, it was impossible for her to be unaware of 
the streak of intense ambition running through her young brother. 
In a moment of prescient candour, he once confided in her: 'I 
won't be satisfied until I come home and land at Heathrow and 
there's a big board with the Amstrad name on it. I want the 
Amstrad name to be as well known as Hoover.' 



49 



The Truck Driver and 
His Wife 



Alan Sugar was scratching his brain about how to get Amstrad 
into Comet. There could be no more important breakthrough for 
his business in the early 1970s than to have his goods on the 
shelves of Comet's stores. By then he had launched a string of 
products for the fast-growing audio industry - parts like amplifiers 
and tuners which went to make up a hi-fi system - and he knew 
that Amstrad and Comet were made for each other. 

In its time Comet was one of the true retailing revolutionaries. 
It had the same transforming effect on the sale of audio and hi-fi 
equipment as Habitat later had on household furnishings and 
Next on men's clothes. Like all retailing pioneers, Comet's success 
flowed from its recognition that the aspirations of a new group of 
consumers were not being met. And it so happened that Comet's 
and Amstrad's natural constituencies were identical. 

Comet's roots lie in a small family business in Hull. Michael 
Hollingbery, who led the company's discounting revolution, 
inherited just one shop from his father. His key decision was made 
in 1968 when he began to offer electrical goods for sale below the 
prices recommended by the manufacturers. The manufacturers 
loathed the idea of discounting, but they could do nothing about 
it thanks to the Government's abolition in 1964 of the rigid set of 
pricing rules known as retail price maintenance. 

A whole new philosophy of selling audio equipment accom- 
panied Comet's decision to discount. Comet opened large stores 



50 



ALAN sugar: the amstrad story 

on the edge of town which were notable for their absence of frills, 
and carried this spartan philosophy into its advertising. A Comet 
advert was little more than a list of the equipment that it held in 
stock, identified by the manufacturers' names and product descrip- 
tions, together with the discounted prices. A customer could buy 
the goods either through mail order or by visiting a Comet store. 

Before the emergence of Comet, the audio market had been 
dominated by small independent outlets, made for hi-fi buffs to 
linger lovingly over the latest breakthrough in high quality sound. 
Comet and Hollingbery were angling for an entirely different 
market: newly affluent young consumers, who wanted their own 
hi-fi set because they had been reared on pop music but did not 
want to bother about the technical details of the equipment. They 
wanted something which worked, looked good, gave value for 
money and could be bought with the minimum of fuss. The truck 
driver and his wife' - this was how Sugar was later to sum up 
this market. It was precisely the market which he coveted for 
himself. 

But there was a problem: how could Amstrad make the initial 
breakthrough? Comet's stores had spread rapidly in the north 
during the late 1960s and were beginning to colonize the south by 
the early 1970s. Already keenly aware of its market power, Comet 
knew it could make or break a budding young manufacturer like 
Sugar. With plenty of other fish in the water, there was no reason 
for it to favour Amstrad. 

Sugar fretted long and hard about Comet. The Sugars spent a 
restless holiday in Majorca with the Amstrad boss brooding by 
the poolside about this dilemma. He knew that the first step was 
to secure a face-to-face interview with Gerry Mason, Comet's 
marketing director. 

Mason was a volatile Yorkshireman who made most of Comet's 
key purchasing decisions. Years later he was to affect the life of a 
country squire, ensconced in a newly built 'manor' on top of a hill 
south of Harrogate, with half a dozen retrievers, as many cats, 
stables, a younger wife and a huge leather-topped desk in the 
middle of a study that looked totally unused. Short and stocky, his 
shirt open at the neck to reveal a tuft of white hair and a gold 



51 



DAVID THOMAS 

medallion, with gold sovereign rings on his fingers and green 
wellies on his feet, there hung about Mason even in his retirement 
echoes of the hard-living and hard-swearing reputation which had 
been a legend in the industry. 

Sugar started phoning Mason to pester him for a meeting, at 
first to no effect. 'Eventually, one day I got an interview with him. 
It was about as easy as getting an interview with the Queen,' 
Sugar recalls. 

Mason made it plain that he was not interested in Amstrad's 
new line of amplifiers, precisely the goods which Sugar wanted 
Comet to take. Instead, Mason wanted Sugar's help in getting 
hold of record players sold by Garrard, a leading British audio 
manufacturer of the time. 

Garrard had an odd pricing policy for its record players, which 
it sold as chassis without plinths and covers. It set one price for a 
retail outlet which planned to sell them direct to the general public 
and a lower price for a manufacturer like Amstrad intending to 
incorporate them into its own products to sell as completed units. 

This two-handed pricing policy was almost calculated to open 
up a secondary market, known in the trade as a 'grey' market. A 
canny manufacturer like Amstrad could buy Garrard's record 
players at the cheaper price and then sell them straight on to a 
retailer, allowing Amstrad and the retailer to split the saving they 
had made at Garrard's expense. Sugar explains: 'There was a 
good grey market business going on there. We were a manufac- 
turer, so Garrard would sell us the record players at the cheaper 
price. We could then fit them into our plinths and covers and ship 
them on to the likes of Comet who could sell them as fully made 
up record players.' 

Sugar travelled up to Comet's Hull office to see Mason on the 
understanding that he was going to sell Mason lots of Garrard 
record players at the grey market price. Once safely over the 
threshold of Mason's office, he quickly agreed to the Garrard deal 
— but on one condition: 'I'm only going to do it on the basis of you 
taking some of my amplifiers,' he told the Yorkshireman. 

Mason was reluctant because Comet had no need of another 
line of amplifiers, but in the end the two reached a compromise. 



•52 



ALAN sugar: the amstrad story 

Mason would not buy any of Amstrad's amplifiers, but he would 
include them in the Comet listings. Then if anybody asked for the 
amplifiers, Comet would order them ad hoc. 

Sugar returned from Hull well pleased with the day's business. 
True, Comet had not actually bought any of his gear, but featuring 
the Amstrad name in Comet's listings was the next best thing. On 
the journey back to London, Sugar was already turning over in his 
mind how he could transform this opening into hard orders. 

As soon as Comet published the advertisements with Amstrad's 
equipment in its listings, Sugar moved into action. Mobilizing half 
a dozen relatives, friends and colleagues to write to Comet with an 
order for an Amstrad amplifier, he was careful to ensure a 
geographical spread. Johnny Simons, his father-in-law, wrote in 
from Redbridge, in north-east London. George Chinchen, who 
had designed Amstrad's first amplifier, ordered one from Croydon, 
to the south of the capital. 

Sugar sat back and waited. A week or so later, a clerk phoned 
Amstrad from Comet's head office: Tt appears we've received 
seven orders for your new amplifier. Just to be on the safe side, 
we've decided to take ten.' 

Sugar immediately went into his angry act: 'Don't be so bloody 
stupid. You're supposed to be Comet. I'm not selling you ten of 
anything. You've got to take at least 100 or it's ridiculous. That's 
it,' he shouted down the mouthpiece and hung up. 

Then he waited by the phone, wondering if he had overplayed 
his hand. To his relief, the phone rang again a few minutes later. 
'All right. Send us 100,' said the man from Comet. 

The relationship between Amstrad and Comet quickly flour- 
ished. 'We've never looked back. We've done mega-millions of 
pounds of business with them. But that's how it all started. That's 
how I broke into Comet. I conned them into it really,' Sugar says, 
and grins. 



Alan Sugar had started making amplifiers in 1970, a couple of 
years before his breakthrough into Comet. He realized that the big 
manufacturers would sooner or later catch up with his production 



•53 



DAVID THOMAS 

methods for the plastic dust covers, wiping out his price advantage 
in the process. It was imperative for him to launch a new line of 
products before that happened. 

Early in his back-of-the-van days, Sugar had met an engineer 
called George Chinchen, who hailed from the West Country. He 
helped Sugar out by repairing the dozens of dud radios that the 
latter acquired in his buying and selling operations. Every week, 
Chinchen would come round to Amstrad's St John Street premises, 
pick up a few radios, fix them for half-a-crown (12V2p) a set and 
then Sugar would re-sell them. 

His success with the plinths and covers set Sugar thinking about 
the other bits which went to make up a hi-fi set. He had been 
impressed by Chinchen's claims to understand the insides of audio 
equipment, so he put two-and-two together and made him a 
proposal: 'O.K., George, how about making an amplifier? I'll 
finance it and we'll go into production.' 

It was not long before Sugar was hawking Amstrad's first 
electrical product, the Amstrad 8000 stereo amplifier. St John 
Street was not big enough for a full production line, so he looked 
around for larger premises. In 1971 Amstrad moved into an area 
of small warehouses and workshops at the back of Fleet Street and 
within a few hundred yards of St Paul's Cathedral. His new 
premises in a four-storey, wide-fronted building in Great Sutton 
Street offered Sugar more space than his first headquarters, and 
he was able to install his first small production line with the help 
of some associates of George Chinchen. Amstrad was soon employ- 
ing a couple of dozen people, mainly women assembling bought- 
in components to make the Amstrad equipment. 

Looking back, Sugar has no illusions about what he was selling 
at first: 'It was the biggest load of rubbish I've ever seen in my 
life. Absolute garbage, it was. I didn't know any better.' No sooner 
was the 8000 amplifier on the streets than he began to think about 
an improved version. 'It was trial and error, because I didn't 
know anything about amplifiers or hi-fi. I used to read the 
magazines, trying to understand what people wanted. I read about 
watts and frequency response, and all that sort of stuff.' 

Sugar would discuss with Chinchen proposals for an improved 



54 



ALAN sugar: the amstrad story 

model, quizzing the other man on how well a prototype amplified 
a musical note: 'Is that good frequency response, George?' 
Chinchen would invariably reply: 'Well, not really, you'll have to 
spend more money on it.' 

So Sugar duly launched the Amstrad Mark II stereo amplifier 
in 1971. Its great improvement over the earlier version was a filter 
to reduce noises from scratches on a record - designed 'to further 
enhance its remarkable and outstanding performance' as 
Amstrad's adverts put it at the time; or 'another marketing ploy' 
as Sugar describes it now. 

Sugar's homework among the hi-fi magazines shines through 
his adverts for the Amstrad 8000 Mark II. The readers of Hi-Fi 
NewS) a leading journal for the buffs, were told in a December 1971 
advert that soundings 'on our test bench' had demonstrated 'the 
rich harmonics and good transient and frequency response so 
necessary for reproduction'. 

That was not quite how retailers perceived Amstrad's early 
efforts in audio. Malcolm Cross was one of the first to sell the 
amplifiers from his shop in Hackney: 'As many as we used to sell, 
we used to send back. The cabinets were always breaking. We 
would take a whole van-load of repairs to him and he was set up 
to receive them. He'd have them repaired and brought back the 
next day. He had to do it like that, because so many went wrong.' 

But Amstrad's earliest amplifiers had an overwhelming advan- 
tage which more than compensated for the initial quality prob- 
lems: price. The Amstrad 8000 cost just £17.70. Even the Mark 
II, with its 'exceptional flexible facilities', sold for only £27.95, 
undercutting almost everything else on the market. 

Malcolm Cross soon found that he needed only minimal effort 
to sell Amstrad's amplifiers. Youngsters would travel for miles to 
buy Sugar's equipment. They were ordinary working-class lads 
who wanted a sound system in their bedrooms, but could not 
afford an up-market stereo system. Cross began to sell Amstrad 
equipment by mail order and money flooded in from around the 
country. 

When Cross had a spare moment in his shop, he would compare 
the sound quality of the expensive machines with that of 



55 



DAVID THOMAS 

Amstrad's: it was as though they were from different planets. But 
this did not bother the teenagers in their anoraks who carted away 
Amstrad's amplifiers. As Sugar explains: There was the snooty 
area of the public that would never buy an Amstrad hi-fi and they 
went out and bought Pioneer or whatever, and they're 5 per cent 
of the market. The other 95 per cent of the market wants something 
that makes a noise and looks good. And they bought our stuff.' 

Meanwhile, Ashley Morris had been going from strength to 
strength with Global Audio. When Sugar sold his stake in the 
firm, Morris agreed to give special prominence to Amstrad's 
equipment in return for a discount. This understanding continued 
as Global expanded into a chain of ten shops. Global always 
featured Amstrad's line when advertising in the hi-fi and music 
press. 'Global Audio Present Their Velvet Sound of 1973' was the 
headline of one typical ad introducing a splurge of equipment, 
nearly all of it from Amstrad. 

Doing an old friend a favour was the least of the reasons why 
Morris kept to the arrangement. Like Cross, he found that 
Amstrad's line sold in large volumes - and it was not difficult to 
understand why. 

All the names in hi-fi, the Quads and the Pioneers of the world, 
would beat their way to the door of Morris's flagship showroom in 
Notting Hill Gate, one of the trendiest hang-outs for the post- 
hippy rock generation in the early 1970s. Their reps would come 
in with wonderful-looking machines and launch into the patter: 
This one's got 5DB distortion level, and it's got this twinkler 
switch, and this star-gazer switch . . .' They would carry on for 
five minutes in a lingo that only a hi-fi enthusiast would under- 
stand, until it came to the bottom line. How much would it cost? 
Typically £50, or £70, or even £100, just for an amplifier. 

Amstrad's approach was somewhat different. Sugar himself - 
never a rep - would turn up with a new product, dump it on 
Morris's desk and say, 'Get a load of that.' He did not need to say 
much about it, because the bottom line spoke for itself: £20, £30, 
£40, whatever it was, Morris knew immediately that there was 
nothing which could begin to touch it for price. 

The established hi-fi manufacturers were hung up on producing 



56- 



ALAN sugar: the amstrad story 

Aston Martins or Rolls Royces of sound, but Sugar understood 
that the era of mass consumption demanded a Mini. Newly 
affluent workers, taking a music machine into their homes for the 
first time, wanted a box that would do the job, looked attractive to 
them and could be made to work without messing around with 
endless wires. 'No one, but no one ever gave Sugar a run for his 
money. He was the only guy who had a little bit of marketing 
foresight to produce crap/ Morris says. 

Except that Amstrad's equipment was not really that bad, once 
Sugar had begun to improve on the unreliability of the earliest 
versions. Above all, the Amstrad line offered excellent value for 
money, as Morris acknowledges: 'Not one person came along in 
the whole time I was in retail and said this one is more reliable 
than Alan Sugar's, it's much more attractive and it's approxi- 
mately the same price or a pound more.' 

The Great Sutton Street premises soon proved too small for 
Sugar. The success of his fledgling manufacturing operation meant 
he needed to hire more assemblers, testers and packers. In 1972, 
he took a couple of floors in 89 Ridley Road, a large building 
which extends for fifty yards down a side road leading into the 
heart of one of the East End's liveliest street markets. Ridley Road 
market caters for all the racial groups that make up the East End 
working class: white, Asian, Afro-Caribbean and Jewish. Halal 
butchers jostle with sellers of yams, sweet potatoes, bagels, pigs' 
tails, jellied eels and cockles. 

No. 89, almost exactly in the middle of the market, is by far the 
most imposing building in the neighbourhood. Stars of David on 
its parapets and on the railings outside suggest past links with 
Hackney's Jewish community. For Sugar, it was like coming 
home. Ridley Road is only a mile or so south of the Northwold 
estate; his mother used to do her shopping there as he was growing 
up, and was occasionally seen about the market by old friends 
until a few years ago. 

Fay Sugar says she first began to realize her son was becoming 
successful 'when he took that place in Ridley'. The family found it 
difficult to grasp that one of their own was employing dozens of 
people in the same streets which their relatives had recently 



57 



DAVID THOMAS 



pounded looking for work. Fay, with daughter Daphne sometimes 
in tow, would hesitate before mounting the stairs to Sugar's new 
offices-cum-factory: 'We used to be amazed, because we used to 
go out shopping, and we would walk in, and we couldn't believe 
it. This was Alan's factory, and he had all these people working 
for him,' Daphne remembers. 

Sugar's progress registered with the trade too. Ronnie Marks of 
Premier Radio first appreciated that Amstrad was beginning to 
rise above the pack of small manufacturers scrabbling around the 
edges of the audio market when Sugar became established in 
Ridley Road. Yet no sooner had he filled every corner of his new 
premises with women working on his audio equipment than he 
began to change tack. Amstrad, the budding manufacturer, grad- 
ually turned itself into Amstrad, the highly efficient sub-contract- 
ing and importing organization, sowing the seeds for the flexibility 
towards sources of supply which was to become a hallmark of the 
Amstrad philosophy in the 1980s. 



Many of Amstrad's adverts in the 1970s carried the proud words 
'British made' wrapped round a Union Jack. No doubt this was 
largely a marketing ploy at a time when it was fashionable for ads 
to feature these patriotic tags, but it also at first reflected Sugar's 
perception of what he was doing: making British amplifiers, tuners 
and other audio equipment out of British components in the heart 
of London. His attitude changed when he began to notice where 
the components came from. 

w We prided ourselves on being British manufacturers. And we 
used to buy components from Plessey, Mullard, IT&T - all the 
main British suppliers,' Sugar recalls. 

But the words 'Made in Japan' on the boxes in which the 
components arrived stuck out like a sore thumb. Sugar quickly 
concluded he could do without the electronics giants as middle 
men: 'We don't need you, mate, thank you very much, to act as 
an agent for us/ he thought to himself 

Sugar found an agent in Japan to buy the components direct for 
Amstrad. Not possessing even a telex then, he communicated by 



58 



ALAN sugar: the amstrad story 

telegram with his new Japanese agent, Shomei Trading. But the 
effort paid off, as Sugar explains: 'We were able to buy components 
at much lower prices than we were buying from the likes of 
Plessey, who were just importing the things and sticking their 
name on them.' 

The time came when Sugar decided he ought to visit the Far 
East himself The volume of components he was importing from 
Japan was such that he wanted to discuss future plans face to face. 
Sugar and his agent met at Tokyo airport, but there was a moment 
of embarrassment when the agent asked whether Sugar had come 
with his father. Age and seniority count for much in Japan and 
the Japanese businessman had never met such a youthful entre- 
preneur. 'He couldn't get to grips with a businessman in his mid- 
20s/ Sugar notes. 

Unknown to the venerable Japanese businessman, however, the 
young man from London spent much of the visit fulfilling a 
lucrative series of orders which more than paid for his first journey 
to Japan. These orders had come Amstrad's way in a rather 
unusual fashion. 

A couple of months before the Japanese trip, a man had phoned 
Sugar to arrange an appointment in Ridley Road. He worked for 
the United Africa Company (UAC), a subsidiary of Unilever, the 
giant Anglo-Dutch corporation which had extensive links with 
Nigeria. It so happened, he told Sugar, that Nigerian civil servants 
had recently received large amounts of back pay and were keen to 
spend their new-found wealth on hi-fi equipment. The UAC would 
like to buy Amstrad equipment for sale in Nigeria. 

It sounded as though a very large order was in the offing, so 
Amstrad's boss was more than disappointed when an order arrived 
for six amplifiers and six loudspeakers. To make matters worse, it 
was accompanied by elaborate instructions on how the equipment 
should be packed for its journey to Nigeria. Sugar took one look at 
these instructions and threw the order in the bin. 

The Amstrad chairman had almost forgotten about the episode 
when the UAC rep phoned up. T'm sorry but I made a big 
mistake with the quantities. I've had a word with my superiors 



•59 



DAVID THOMAS 

and we need 2,000 amplifiers and speakers. But we need them in 
a hurry. We want as many as possible this week.' 

In those days Unilever meant nothing to Sugar, who had no 
particular reason to trust the UAC man. Sugar reflected that he 
was being asked to supply at short notice equipment worth almost 
£200,000. This is it,' he thought. This is the big sting. I'll hand 
over my amplifiers and never see the money.' 

Sugar made his reservations plain and the UAC rep said he 
would put the money into Amstrad's bank account the next day 
by telegraphic transfer. Sugar had not heard of telegraphic trans- 
fer, but the following day his bank rang up to say that the money 
had indeed been deposited. 

When the UAC rep heard that Sugar was about to go to Japan, 
he asked him to keep his eye out for goods that might interest the 
Nigerians. Sugar thought little of this until he arrived in his hotel 
room in Tokyo to find a message from the UAC asking him to 
arrange the urgent shipment of 5,000 radio cassette recorders to 
Nigeria. 

Amstrad's chairman turned to his agent at Shomei Trading to 
find the cassette recorders, but he felt obliged to keep secret the 
price he was charging the UAC. The point was that I was adding 
a 50 per cent mark-up to the price I was paying for the goods in 
Japan. It was unbelievable. But it would have been embarrassing 
if my agent had realized how much I was making out of the 
business.' 

Sugar found freight forwarders in Japan who shipped the 
cassette recorders to Gatwick for transit on to Nigeria. No sooner 
had he despatched them than another message came through from 
London asking for thousands more pieces of equipment. 

The Nigerian gravy train continued to run a little while after 
Sugar had returned to Britain, helped by the fact that the UAC 
man was innocent of the audio world. Only after the Amstrad 
batches of amplifiers and speakers reached Nigeria did he realize 
that the Nigerians would also need record players. 'He really 
hadn't the faintest idea of what he was buying, 5 Sugar says. 

But the Nigerian civil servants' wealth dried up, and with it 
Amstrad's business in that part of Africa. Shortly afterwards, Gulu 



60 



ALAN sugar: the amstrad story 

Lalvani returned from a trip trying to sell his Binatone equipment 
to Nigeria. He had been amazed to find Lagos, the Nigerian 
capital, swamped with Amstrad amplifiers even though Sugar had 
never set foot in the place. 

Sugar's visit to Japan was to be the first of dozens of trips to the 
Far East. Negotiations over the import of components were the 
main reason at first, to be followed later by visits to sub- 
contractors, as he began to have some of his machines made in 
Japan, Taiwan and Korea. But right from the start there was 
another motive: to take a look at the latest equipment pouring out 
of Japanese factories. Sugar would spend a profitable hour or two 
wandering around Akihabara, the district where Tokyo's con- 
sumer electronics shops are clustered, noting the trends in audio 
equipment, picking up hints about how to style his machines, 
imagining to himself how he could improve on what he saw there. 

Japan also houses an array of companies specializing in making 
audio equipment for export to the West, to be sold under the name 
of a European or U.S. company. Western companies typically 
make smaller profit margins on these badging operations, known 
as OEM (original equipment manufacturing) in the trade, than 
on equipment made in their own factories. Yet OEM business has 
two important advantages for the Western company whose name 
ends up on the equipment: it generates profits without the hassle 
of having to manufacture the equipment, and the OEM factories 
can supply goods which complement those made in-house - a 
turntable to go with an amplifier, for example. 

Sugar grasped the potential of these arrangements and, through 
introductions arranged by his agent, he struck up a relationship 
with a Japanese OEM company called Orion which was destined 
to become one of Amstrad's most important suppliers. Wanting to 
sell cassette tape decks to expand his audio range, Sugar turned to 
Orion to supply them. As a result, in 1974 he was able to unveil 
his 6000 stereo cassette deck. This was quickly followed by the 
7000 series which boasted a Dolby noise reduction facility and had 
been 'designed to provide all the necessary features to line up with 
your high fidelity set-up', as the ads put it. The deck was made in 
its entirety by Orion, but carried the Amstrad name. Nick 



61 



DAVID THOMAS 

Lightowler, Gerry Mason's right-hand man at Comet, remembers 
the episode. 'He imported two tape recorders which were pig-ugly, 
but dirt cheap. They were phenomenally successful, because they 
were phenomenally cheap.' 

This experience set Sugar thinking about what he was doing in 
Ridley Road: Tt opened my eyes again. There I was knocking my 
brains out with dozens of people working in Ridley Road, soldering 
and assembling, and I realized that we were making a profit on 
these cassette decks coming through and not really doing anything 
at all. I was just opening a letter of credit and importing the 
things.' 

The answer, of course, was not to transform Amstrad totally 
into a badging operation, because that would have deprived the 
company of the greater profits to be won from controlling the 
design and manufacturing of the bulk of its equipment. But Sugar 
reckoned there was an easier way to make money than spending 
his life supervising the women in Ridley Road. T'm not going to 
knock myself out. I'd rather sub-contract things out,' he 
concluded. 

Not long before Sugar reached this conclusion, Ron Nixon 
bought a chalk-pit just off the A2 between Maidstone and Roch- 
ester, in the middle of the North Kent belt of light engineering 
factories. He and his partner, Len Lawrence, were set on expand- 
ing their business, L. & N., which had flourished in the 1960s 
assembling toys and bits of audio equipment. Now Nixon intended 
to build a new factory on the site of the chalk-pit. 

He struck a deal with Fidelity Radio, then a force to be reckoned 
with in Britain's high streets, whereby L. & N. would make a 
range of Fidelity hi-fis and televisions. But the factory for the 
Fidelity contract would occupy only about half L. & N.'s new site. 
Nixon began casting around for another customer. A close 
observer of the audio market, he had become aware of this 
newcomer, Amstrad, in the previous couple of years. So he tried a 
direct approach. 

Nixon's call to Sugar could hardly have been better timed and 
the two agreed that Amstrad would sub-contract to L. & N. the 
job of making most of its audio equipment. With this deal in his 



•62 



ALAN sugar: the amstrad story 

pocket, L. & N. built a second facility on the chalk-pit dedicated 
to Amstrad work. The foundations of the Amstrad plant were laid 
just as L. & N. was opening its Fidelity factory. 

From 1974 for the rest of the decade, most of Amstrad's 
equipment that was not imported in finished form from the Far 
East flowed through L. & N. At its height the factory had about 
120 people working full-time on Amstrad's requirements. Ampli- 
fiers, tuners, tape decks, compact music centres - the expanding 
range of Amstrad equipment all emanated from L. & N.'s plant. 
Ironically, the official description of Amstrad in its annual report 
and accounts changed from 'wholesaler' to 'manufacturer of 
electronic equipment' in 1973, just as Sugar was preparing to sub- 
contract out most of his manufacturing activities. This was not to 
be the last time that Amstrad's self-image was slow to catch up 
with the company's fast changing character. 

Sugar settled into a pattern with L. & N. during his frequent 
visits to their factory. Early on in the year, he would bring down 
prototypes of the new models for the coming Christmas season. 
He controlled not just the designs, but also component purchasing. 
It was L. & N.'s job to place the components into the printed 
circuit boards, assemble all the parts into the finished products, 
test them, box them with the Union Jack clearly displayed and 
send them off to Ridley Road for final testing and packaging. 
L. & N. began production of the season's models early in the year, 
building up to a climax in the second half as the exact pattern of 
Christmas demand emerged. 

Ron Nixon soon learned to understand - and appreciate - the 
way Sugar operated. 'He was in bold all the time. He'd go round 
and rave and rant at anything that wasn't right,' remembers 
Nixon, a heavy-set, slow-speaking man now into his seventies who 
in his day was not beyond a bit of ranting himself. 'We used to 
have some real old slanging matches.' 

Throughout the 1970s, L. & N. grappled with a chronic short- 
age of skilled workers - one of a gamut of problems which 
eventually persuaded Sugar to switch even more of Amstrad's 
manufacturing to the Far East. Ron Nixon explains its impact: 
'We gave him problems, because there was a tremendous shortage 



63 



DAVID THOMAS 

of people in those days. You couldn't find anybody who knew a 
resistor from a capacitor if you walked from here to Russia. So we 
gave him certain aggravation, because we couldn't get the people 
to do as much as he wanted or to maintain the quality that he 
wanted.' 

Sugar's visits were always conducted at whirlwind speed, as 
Nixon recalls: 'He wouldn't stay long. He wasn't the sort of man 
to say, "Let's go out for some lunch." Or if he was here over lunch, 
it would be a sandwich in a pub. Then he would rush off. He 
would never waste time.' 

The Amstrad boss was always pressing L. & N. to turn out his 
machines more quickly and in greater quantities, but beyond these 
overriding preoccupations Sugar invariably focused on one factor. 
'He was mainly interested in the finish of the thing. Amstrad was 
more fussy cosmetically about how the thing looked than Fidelity,' 
Nixon learned, although Sugar was totally unconcerned about 
technical details: T don't think he cared tuppence how it was 
made, as long as it was made.' 

Sugar could not always control the appearance of equipment 
which he imported from the Far East simply to badge, but from 
his earliest days as a manufacturer of audio equipment, he had 
insisted that equipment designed by Amstrad should look good. 
His advert for the 8000 Mark II amplifier, his second audio 
product, proclaimed: 'All this electronic technology is housed in 
an elegant low-line, decor-styled satin teak finished enclosure with 
matt black facia with silver trim and brushed aluminium controls.' 

Right from the start too, although uninterested in technical 
details, Sugar seized on one measure which he believed would sum 
up the power of his machines in the eyes of his customers - the 
watts they were capable of delivering. Believing technical ignor- 
ance to be the norm, he reckoned that most people going into a 
shop would latch on to one measure of up-to-dateness: they would 
ask about the amplifier's wattage. What mattered was for Amstrad 
to match the industry leaders on this one measure. It was a lesson 
he would apply later to computers. 

For the fact was that while Amstrad equipment might have 
seemed cheap and even vulgar to hi-fi experts, Sugar was not 



64 



ALAN sugar: the amstrad story 

trying to sell to the connoisseur. He knew that people like the 
truck driver and his wife who bought Amstrad's machines did not 
need to hear the second violins in an orchestra. He also knew that 
an expensive-looking finish was one of the main features that 
would attract the working-class and lower-middle-class purchasers 
of his equipment. He did not need to rely on market research, 
consultants or a textbook to tell him this; he understood their 
needs and their aspirations because he was one of them. 



The full glare of publicity was turned on Amstrad's years of 
spectacular growth in the 1980s, when the company launched into 
computers and word processors. Yet Alan Sugar had been there 
before. 

In the 1970s, he had to cope with rates of growth that ranked 
with anything he experienced a decade later, though obviously 
from a much lower base. In 1971, when the first line of amplifiers 
began to take off and he was still making the plastic dust covers, 
sales more than doubled to £207,534 and profits rocketed eleven- 
fold to £24,242. The next year saw a further tripling of sales and 
profits. In 1973, when Amstrad began to export to France, the 
figures doubled again, with the company notching up profits of 
£194,063 on sales of £1,327,366. 

Coping with that growth meant he had to start hiring managers 
to whom he could delegate some of the work. Sugar signed up an 
engineer and a draughtsman to help design the audio equipment, 
a sales manager to link with his customers and a purchasing 
manager to deal with suppliers. Amstrad was no longer quite a 
one-man band, although all the company's activities still revolved 
around that one man. 

Sugar's relations with his early customers, the dozens of inde- 
pendent audio retailers, began to take on a new light. The Amstrad 
boss had always been straightforward to the point of bluntness 
with his customers. Secure in the increasing volumes of his goods 
that were passing through Comet, he could now choose more 
carefully among the small independent retailers who as a group 
still dominated audio retailing. He could afford to drop those who 



65 



DAVID THOMAS 

took a long time to pay, who ordered two pieces of equipment and 
sent back one, the incompetent or fickle dealers who were destined 
to stay in the small league. 

Selling through the independents meant dealing with hundreds 
of outlets stuck in crowded streets in the centre of towns. By 
contrast, Comet offered Sugar economies of scale that pushed 
Amstrad's profits into a new league. Comet was in the business of 
shifting boxes in very great numbers and Amstrad was able to 
drop off large deliveries at the back of each Comet store, perfectly 
located near fast out-of-town roads. 

Amstrad's links with Comet became closer throughout the 1970s 
when Sugar was a frequent visitor to the offices of Gerry Mason 
and Nick Lightowler. He would make the trek to Leeds or Hull, 
where Comet's operations were based, about half a dozen times a 
year. Sometimes Sugar would be there to negotiate what he called 
a 'blow your brains out' operation, when he wanted to clear out a 
couple of thousand pieces of an old line in a hurry by slashing 
prices. His most important visits were made in the late spring and 
early summer, the period of peak preparation for the Christmas 
season, when he would discuss with Comet what he was planning 
for the season and agree with Mason or Lightowler the terms of 
the deliveries. 

Mason had come across many brash young entrepreneurs like 
Sugar trying to make their way in the fast-growing audio world. 
While some showed early promise, in the end most fell down in 
some respect: the quality of their equipment was too poor, or the 
price was not right, or they simply could not deliver when the 
goods were needed. 'When it was quiet, you'd get all the supplies 
you needed at the right price, but when it was busy, you wouldn't 
see them,' Mason says. 

Not so with Amstrad. Comet found that Sugar kept to delivery 
dates, a quality particularly unusual at the lower end of the market 
which was heavily dependent on the often unreliable flows of 
imports from the Far East. Sugar built up an enviable reputation 
with Comet for keen pricing within his segment of the market, as 
Mason notes: 'If Alan was selling it, you wouldn't buy it cheaper 
from anyone else.' 



♦66 



ALAN sugar: the amstrad story 

Mason was the first executive in a large company to experience 
what at first sight seems a surprising side of Sugar's character. He 
found Sugar easy to deal with. Easy, it must be stressed, does not 
mean weak or soft. On the contrary, negotiations between Amstrad 
and Cornet were often fierce on detail and aggressive in style. 
Mason met Sugar head on: 'Alan used to curse like an old trooper. 
But then I would curse back at him too.' 

When Mason says Sugar was easy to deal with - and it is a 
comment echoed by many, if not all, of Sugar's later business 
associates - he means that negotiating with him was straight- 
forward, that Mason knew where he stood. The Amstrad boss 
would set out his stall, telling Comet precisely what products he 
could deliver over what period, together with the prices he would 
charge and the credit arrangements he would give Comet. Sugar 
would then listen to any alternative ideas that Comet had, show 
surprising flexibility in amending his initial plans where he could 
and make it perfectly clear where he would not meet Comet's 
counter-proposals. 

Relations between suppliers and retailers in cut-throat areas of 
the high street are fraught with possibilities of two-timing. A 
supplier will swear blind to a particular retailer that he is getting 
better terms than any of his competitors, and promptly say the 
same thing to the next retailer. Mason was used to dealing with 
sweet-talking manufacturers and then finding out through the 
grapevine that they had been economical with the truth. But he 
says of Sugar: 'When Alan told me that Comet was getting the 
best price from Amstrad, I never found him out in a lie. And he 
never welshed on a deal.' 

Nick Lightowler, who continued to have an important role in 
Amstrad's fortunes in the 1980s when he became purchasing 
director of Currys in the Dixons group, was impressed by Sugar's 
reaction to any new proposals from Comet. Sugar's first response 
would usually be: 'Naw, it can't be done.' Then invariably he 
would phone Lightowler a few r days later: 'You know that thing 
we were talking about, well, I've been thinking about it and this is 
what I've come up with . . .' From such incidents, Lightowler 



67 



DAVID THOMAS 

formed a lasting impression of Sugar as 'a guy who would work on 
how it could be done, rather than on why it couldn't'. 

There was another major advantage about negotiating with 
Sugar. As far as Comet was concerned, Sugar was chairman, 
managing director, sales director, finance chief, distribution man- 
ager, technical supremo and sales representative all rolled into 
one: no one else in Amstrad ever got near the Comet deals. Mason 
found dealing with other companies 'more of a long negotiating 
job'. The pattern with Japanese groups, with Philips or even with 
most of the British audio companies was for Comet to discuss 
business regularly with one contact point, typically the sales or 
area manager, and then to negotiate a main contract with the 
chairman or managing director once a year. This meant that 
Comet's regular contact had to refer back most decisions during 
the year to head office. Sugar, by contrast, was able to make 
instant decisions about any aspect of his business, a freedom 
perfectly suited to his temperament. 

As Amstrad grew hand-in-hand with Comet during the 1970s, 
Sugar and Mason often found themselves in each other's company 
during visits to American audio fairs in Chicago or Las Vegas, 
where they would scour the exhibits for ideas as to what might 
catch on in Britain. On these trips away from the minute-by- 
minute pressures of running the business, Sugar unwound: 'He 
was always one of the boys. If we were going out on the booze, 
Alan would come along with us. If there was a trip to the Grand 
Canyon, he was always there,' Mason recalls. 

Yet the process of relaxation went only so far; when the Comet 
director and the Amstrad boss found themselves in some bar or 
restaurant thousands of miles from home, their conversation 
usually turned to business: 'We never stopped talking about 
business in those days. It was all we used to talk about. We both 
loved what we were doing,' says Mason. 

Both then and since, Sugar's associates have always found his 
focus intense when he turns his mind to something. But he never 
bothered to focus on the keepers of conventional wisdom in the 
established hi-fi industry as epitomized by the trade press. His 
tactic was to ignore them and, if that became impossible, to treat 



•68 



ALAN sugar: the amstrad story 

them with disdain. 'What we were doing was making money, 
while what the others were doing was getting rave reviews in 
magazines that only a tiny percentage of the market was buying,' 
Sugar explains. 

Hi-fi specialists repaid Sugar's contempt - at first by indiffer- 
ence, then through hostility to Amstrad. Arthur Ord-Hume, a 
leading audio journalist for twenty years and later editor of 
Independent Electrical Retailer magazine, recalls the complex of 
prejudices which greeted Sugar: 

His operation was quite frankly rather looked down upon by the 
serious side of the industry. He had all the appearance and the 
trappings of the back-street marketeer, while the attitude of the 
trade press was to support the activities of those who were 
thought of as 'legitimate' traders. Beards weren't very popular 
in those days and he always looked slightly dishevelled. He was 
emerging at the time of the aspiring youngster and it was 
difficult for the industry's establishment to understand or accept 
this. 

The first reviewer in Hi-Fi News to cast his eye over a piece of 
Amstrad equipment - the IC2000 amplifier reviewed in April 1972 
- ran through a string of grouses about the tone control and the 
filters and then exploded with the complaint: T cannot help feeling 
that Amstrad set out to provide too many functions for too little 
cost'. (This must rank as the earliest definition in print of a core 
component of the Amstrad philosophy.) Dimly aware that maybe 
he was missing the point, the reviewer rounded off with this 
declaration: 'Admittedly, manufacturers have to design with the 
market requirements in mind, but we don't have to review in that 
vein!' 

This tone of aggrieved incomprehension was to recur in the very 
few reviews of Amstrad equipment which crept into the pages of 
Hi-Fi News in the 1970s. In a comparison of the Amstrad 5050 
receiver with competitive offerings from JVC and Toshiba in June 
1977, the reviewer surpassed himself by declaring: 'The Amstrad 
is not bad looking by any means; but detailed exploration of its 



69 



DAVID THOMAS 

nooks and crannies leaves one feeling a trifle sad' - a cry from the 
heart which no doubt kept Alan Sugar awake at night. 

Yet between 1972 and 1977, when these two reviews were 
published, many of the industry's pundits had been forced to 
revise their opinion of Amstrad, as Arthur Ord-Hume describes: 

Suddenly, we were aware of the fact that this back-street trader 
was one of the most significant people in the industry. Once he 
began to get distribution in places like Comet and Rumbelows, 
he began to make inroads into the markets of the 'legitimate' 
suppliers. When we began to find Amstrad equipment alongside 
that of Philips and Sony, we had to sit up and take notice. 

Some specialists understood that Amstrad was about value for 
money. A reviewer in Hi-Fi News of Amstrad's 3000 Mark II tuner 
in February 1976 - having noted that it was difficult to find a 
competitor's product for less than £100 - confessed that he found 
it almost difficult to believe that the Amstrad price tag of £60 
included 25 per cent VAT. Even the journalist who was overcome 
by melancholia when exploring the nooks and crannies of the 
Amstrad receiver acknowledged that, compared with the Toshiba 
and JVC machines, Tt is the least costly of the three and sports 
the most features.' 

There were some commentators in the mid-1970s who also 
recognized that the quality of Amstrad's equipment was not at all 
bad — for its price range. Some of Amstrad's earliest amplifiers 
had been of poor quality, assembled out of the barest number of 
the cheapest quality components available. If one component 
failed, it was often less hassle to throw away the piece of equipment 
than to try to repair it. But as Sugar gained experience of the 
business, he steadily improved the reliability of his machines. 

True, Amstrad was not beyond the odd trick. Robert Lawrence, 
son of one of the partners in L. & N., was a young man working 
on the production line when Amstrad sent down the specifications 
for its EX range of tuners in 1976. He noticed that when he 
pressed a switch designed to measure the quality of the sound, the 
needle always moved to the dead centre, indicating perfect sound. 



70 



ALAN sugar: the amstrad story 

An examination of the machine's workings showed them to be 
arranged so that the needle would automatically do this. 'When 
you pressed the switch, it cut off the meter, so that the needle 
automatically fell in the middle. It just got completely switched 
out of the circuit,' Lawrence recalls with a laugh. 

Later, when Amstrad started doing business with Rumbelows, 
the chain's service department opened up an Amstrad speaker and 
found that one of the three cones in the speaker boxes was not 
connected; it seemed to be there just for show. Sugar continued to 
insist in the face of some scepticism that the cone had a purpose in 
modifying the sound; but the incident quickly entered the Amstrad 
mythology, and everyone in the know at Rumbelows joked about 
it for months. 

Yet such tales disguised the more mundane story of steady 
quality improvement. Ron Nixon at L. & N. was able to compare 
Amstrad's attitude to quality with that of Fidelity, then a better- 
known brand. He says Sugar was the more obssessive of the two 
about reliability. He did not care what was inside the machines, 
but he wanted them to work and to look good. 

As Amstrad's biggest customer by far, Comet was in an excellent 
position to assess the quality of Amstrad's products. Both Mason 
and Lightowler insist that Amstrad's fault rate was not out of line 
with other equipment in the same price range, a view shared by 
David Johnson, then managing director of Rumbelows. Mason 
explains: 'He wanted to make cheap and cheerful, but not cheap and 
nasty. In its bracket, it was as good as anything you could buy.' 

Three factors conspired to convey the opposite impression, 
however. First, memories of the poor quality of Amstrad's earliest 
machines lingered. Second, opinion formers in the industry, the 
hi-fi pundits, would not have dreamt of having a low end machine 
like an Amstrad in their own homes, and so failed to appreciate 
that Amstrad's equipment performed adequately for its price range. 
Third, because Amstrad sold more machines than its rivals, the 
absolute number of Amstrad boxes returned for repair was higher, 
even though the relative number was typically the same or lower: 
many shop assistants did not appreciate this distinction and by 
word-of-mouth fed Amstrad's reputation for unreliability. 



71 



DAVID THOMAS 

This reputation still stuck to the company much later when it 
entered the computer business. It was to he many years before 
most Amstrad products were generally recognized as being excel- 
lent value for their niche in the market. 

By shunning the trade press and the rest of the industry 
establishment, Sugar was following a sound commercial instinct. 
He had nothing to gain and arguably much to lose by trying to 
compete on the same basis as up-market hi-fi equipment. Yet by 
refusing to submit to the industry's judgement he was also 
fulfilling a more basic need. As a man who in his mid-twenties 
had forced himself without help or favours into the heart of a 
tightly meshed industry, the one thing Sugar loathed was to be 
underestimated. 

Sugar's aggressive reputation among suppliers and customers 
not only reflected his natural character, but also acted as a defence 
against what to him seemed the ever-present danger of being 
patronized. In a passing conversation about a particular parts 
supplier, he once confessed to Ronnie Marks of Premier Radio, 'I 
can't stand doing business with him because he treats me like a 
little schoolboy. He always calls me "son". I'd rather talk to his 
storeman.' 



As the 1970s progressed, Amstrad's success was flowing through 
into Sugar's standard of living and he was beginning to live the 
life of a successful small businessman. In 1972 the Sugars moved 
to a house in Chigwell Rise, a long road on the borders between 
Essex and Redbridge which is home to many former East Enders 
who have made it. Many of the driveways of the detached houses 
in Chigwell Rise boast a Jaguar or Mercedes. Although the Sugars' 
new home was by no means the most opulent in the road, the fact 
that they could live there at all spoke volumes about Sugar's 
progress. 

Bit by bit, he was acquiring the trappings of wealth. By the end 
of the 1970s Amstrad had provided him with a Rolls Royce as a 
company car, a fitting accolade for someone whose prime goal ten 
years previously had been 'wheels'. Sugar took flying lessons at a 



72 



ALAN sugar: the amstrad story 

local airport, pitching himself into his hobby with typical single- 
mindedness. He passed advanced all-weather tests as well as his 
ordinary pilot's licence. 

In 1975 Amstrad duly bought a plane for £7,500, allowing Sugar 
to make the trips down to L. & N. by plane. He cut two hours off 
the journey by landing his small aircraft on the grass runway at 
Rochester airport. He also flew to his meetings with Comet in Hull 
- dodging around the large chimney at the end of the runway at 
Brough airport as he came down out of the bank of mist lying low 
over the Humber. 

One day Sugar had a lucky escape when flying his single- 
engined aircraft back from France. He had noticed something a 
bit odd with the machine when it was over the Channel, but 
landed as normal at Southend airport to clear customs. After 
taking off from Southend, the real problem started, as he recalls: 
The engine crocked when I was hovering over Southend town. So 
the only way out was to turn right towards some fields and put the 
thing down there.' 

The shattered cylinder which caused the problem still sits in 
Sugar's office as a memento. He decided that prudence required 
his next aircraft to be twin-engined, but twin-engined planes bored 
him: Tt really took the fun out of flying. The single-engined plane 
had been simple: I just got in it, took it up, went round in circles 
and looked out the windows. A twin was too serious a form of 
flying. It's not for just flipping around in.' 

So Sugar went back to a single-engined machine, much against 
the advice of his wife, Ann, whose initial nervousness about her 
husband's new hobby was reinforced by the Southend crash. In 
the early 1980s, reckoning that he had an additional responsibility 
as the chairman of a public company, Sugar bowed to Ann's pleas 
to quit flying. By then, he had discovered a new hobby into which 
he could channel his energy: tennis. 

Flying became available to Sugar as a hobby thanks to the 
wealth flowing to him from Amstrad's success, yet in a more 
important sense he was little changed by this success. True, he 
lost touch with some of his old friends after he moved to Chigwell 
Rise; but the new friends he made were of the same background 



73 



DAVID THOMAS 



and with the same tastes. Neither in the 1970s nor in the 1980s, 
when he could have chosen to move among the wealthiest and 
most successful in the land, did he fundamentally alter his way of 
life. Indeed, many of the friends he made in the early 1970s who 
have remained close to him were part of the extended Sugar- 
Simons family. 

Shortly after establishing themselves in Chigwell Rise, Alan and 
Ann Sugar took to visiting Gerry and Norma Eriera, who lived in 
an even smarter house just down the road. Norma was Ann's first 
cousin, but the Erieras were 8-10 years older than the Sugars, an 
age gap which had kept them apart until they became near 
neighbours. 

Naturally more extrovert than the Sugars, the Erieras are a 
friendly couple who laugh a lot and know how to enjoy themselves 
on an evening out. Alan Sugar found that he could unwind in their 
company, and pretty soon the two families were arranging to 
spend their holidays together. 

Early on Sugar acquired a taste for relaxing through expensive 
holidays. For instance, the Sugars, Crosses and a group of other 
friends spent New Year's Eve 1972 at the Paris Hilton, having to 
abandon their first-class plane tickets and return by hovercraft 
when Paris became fogbound. Over the next decade, outings like 
that were to become a regular part of life for the Sugars and the 
Erieras. 

Alan Sugar put his business skills at the service of his new 
friend. A businessman who has spent all his life in the ladies' 
garment trade, Eriera had never tapped the rich seam of clothes 
made in the Far East before meeting Sugar. He accompanied 
Sugar on one of his trips to Hong Kong and South Korea, where 
the latter showed an unsuspected talent for the garment business. 
Eriera sat quietly next to Sugar in a raincoat factory while the 
Amstrad chairman bartered over raincoats as though they were 
amplifiers. 

Sugar turned to Eriera and asked, 'Well, how many could you 
sell?' 

Eriera replied that he was not sure. After some more haggling, 
Sugar said, 'We'll get a good price if we buy 15,000.' 



74 



ALAN sugar: the amstrad story 

So Sugar duly bought 15,000 raincoats on behalf of his friend 
and also put some money into Eriera's business, allowing him to 
build up his trade with the Far East. Back in Britain, he would fly 
Eriera over to Le Touquet for lunch on a Saturday. Twice a year, 
the Sugars and the Erieras would spend a couple of weeks on 
holiday together. The Grand Hotel at Rimini was the venue of 
their first family holiday, followed by jaunts to Hawaii, San 
Francisco, Bermuda, Fort Lauderdale and a dozen other far-flung 
places. The Erieras introduced the Sugars to Los Monteros, a 
hotel which boasts ten tennis courts, its own golf course, a riding 
club and a private beach, and where suites can cost more than 
£300 a day. The hotel in Marbella, one of the main haunts of 
Europe's super-rich, quickly became a favourite with the Sugars, 
who still regularly spend a fortnight there. 

Alan Sugar was able to forget the pressures of his business 
totally while on holiday with the Erieras. He could indulge his 
growing enthusiasm for tennis and his passion for flying. Staying 
in the Catskill Mountains, he flew with an instructor to Albany 
airport for the afternoon to pick up his American pilot's licence. 
Another time in Hawaii, when the Erieras were on the golf course 
Sugar amused himself by flying over the island and buzzing their 
hotel. 

These pastimes and holidays were way beyond the reach of the 
vast majority of British people. Yet it was possible to say without 
affectation, as Sugar did in 1980 at the time of Amstrad's flotation, 
that at heart he was still one of them. Asked what he spent his 
money on, he answered: 'Oh, clothes and gadgets and trinkets and 
toys, video cameras, calculators that sing songs, spelling machines 
for the kids.' 

Despite Amstrad's growing success, Alan Sugar still went to the 
office every day at 8.30 am, came home at 6.00 pm, watched 
television, read three books a year, played tennis at weekends, 
spent as much time as he could with his family and saw the same 
friends he had always seen. 'I know where my roots are,' he 
explained. 'My family are still working-class people.' In the 1980s, 
Sugar was to become one of that rare breed: a working class multi- 
millionaire. 



75 



• -A • 



5 
A Mug's Eyeful 



Alan Sugar came back from a trip to the Far East in 1978 with the 
germ of an idea which turned into the first blockbuster Amstrad 
product. In the second half of the 1970s, Japanese audio companies 
had begun to group the separate elements of a hi-fi system - the 
amplifier, cassette deck, tuner and record player - into one rack. 
Sugar was among the first in the British industry to appreciate the 
significance of the move: it allowed a manufacturer to sell all the 
hi-fi pieces to a customer at the same time. But it was his own 
trick to go one step further. 

The Amstrad chairman pondered what the racking development 
meant to consumers. They benefited by being able to buy all the 
hi-fi pieces in a common design in one place; yet once they had 
taken the parts home, they still had to go through the tedious 
process of linking them together, which meant grappling with a 
whole mess of wiring before they could use their system. A typical 
hi-fi system looked as though it had a fistful of spaghetti hanging 
out of its back, and it might also have up to four or five plugs. 
Sugar realized this made little sense. 

'The wallies that ended up buying this stuff didn't know what 
they were buying anyway. Once they had fitted them all up in 
their houses, connected all the wires from the tuners to the 
amplifiers and the cassette deck, and put the record player on, the 
thing never got moved ever again,' he comments. 

Nor did this make sense from a manufacturing point of view. 



76 



ALAN sugar: the amstrad story 

The separate pieces in a rack system all had to have their own 
components. Thus there were power supplies, transformers and an 
array of other electronics in each part. Sugar's instinct for cutting 
out duplication, one of the most powerful in his make-up, was 
roused. Tt was obvious to me that a lot of this was a waste of 
time,' he recalls. 

His solution was simplicity itself. He made a front panel to look 
as though it were three separate units - amplifier, tuner and 
cassette deck - but in fact it was a single box. Sugar christened his 
idea the Tower System. His one-plug hi-fi was a breakthrough for 
people who had no taste for the wiring gymnastics demanded by 
conventional hi-fi units. By eliminating duplicated components, it 
also allowed a significant advance in cost reduction. Anyone 
opening up a Tower System would have been amazed to find very 
little there — one set of electronics and a hole. 

Sugar was careful not to stint on the appearance of the front 
panel. Make a cheap product look expensive: this cardinal prin- 
ciple of the Amstrad philosophy had already served him well. The 
silver-coated front of the Tower System was festooned with 
flashing lights, knobs and buttons. It passed the Amstrad test: it 
was, Sugar declared, 'a mug's eyeful'. 

By 1979, Sugar knew that he was sitting on a winner with his 
Tower System and realized that it was the first Amstrad product 
likely to sell in large volumes in the high street. Although he had 
sold some amplifiers and other lines through chains like 
Rumbelows, few of Amstrad's previous products were of a kind 
to take the high street by storm. The Tower System changed 
that. 

Breaking into this area of the market in a big way offered Sugar 
not only larger volumes, but also the chance to reduce his 
dependence on Comet. By the end of the 1970s, Comet was 
accounting for over a quarter of Amstrad's sales, leaving Sugar 
exposed to any downturn in Comet's fortunes. He showed the 
Tower System to Woolworths, who were eager to buy it. 

First the Amstrad boss had to broach his plans with Gerry 
Mason at Comet. The firm was naturally suspicious of a move by 
Amstrad into the high street. If Amstrad built up other customers 



77 



DAVID THOMAS 

to rank with Comet in importance, then Comet's bargaining power 
with Sugar would be correspondingly reduced. More important 
still, Sugar would inevitably become more reluctant to allow 
Comet to discount Amstrad's goods for fear of offending the high 
street chains. 'Over my dead body 5 was how Mason responded 
when he heard about the plans. It was not a threat which worried 
Sugar. 

The Tower System was set to become Amstrad's most important 
product as the 1970s came to an end. Rumbelows and Currys 
bought the machines once Woolworths had taken the plunge by 
becoming the first high street chain to stock them. Comet also sold 
them, after Mason realized he could not hold Sugar back. David 
Johnson, then managing director of Rumbelows, was at one stage 
selling up to 1,000 a w^eek from his chain alone. In the early 1980s, 
after Amstrad's flotation, the Tower System became the first 
Amstrad product whose sales were measured in the hundreds of 
thousands. 'That was when it really started booming, when we 
began to understand what orders were all about. People like 
Woolworths were banging in orders for 5,000 at a time. They sold 
for £199 in the shops, which meant £130 each for us - that was a 
lot of money,' says Sugar. 

One group unimpressed by the Tower System were the hi-fi 
buffs, the protectors of the industry's virtue. 'Anyone who thought 
from the technical point of view frowned upon it. They thought 
what we had done was disgusting. To them, the idea of hi-fi was 
that you had to separate all the parts,' Sugar notes. 

Their reservations cut little ice with him. 'We were attracting 
the truck driver and his wife. They saw this bloody big mug's 
eyeful in this cabinet. This whacking great big cabinet, with a 
beautiful front panel with knobs on it. It was a very uncomplicated 
product. Plug it in the wall and away you go.' 



Early in 1979 Alan Sugar was approached with a proposition. 
Would he sell Amstrad? The unexpected proposal was made by 
Audiotronic Holdings, a quoted company which had swallowed 
up two of the most successful independent hi-fi retailers, Lasky's 



78 



ALAN sugar: the amstrad story 

and G. W. Smith. A man called Geoffrey Rose, who enjoyed a 
brief reputation as a company doctor at the end of the 1970s, had 
been brought into Audiotronic in 1978 to revive its flagging 
fortunes. Rose reckoned that one way to turn Audiotronic around 
would be through acquisitions and his eye lit on Amstrad, by then 
a force in the audio industry. 

The deal Rose had in mind was to pay Sugar about £2 million 
for 75 per cent of Amstrad, leaving him with a 25 per cent stake. 
Sugar would continue to work for Amstrad/Audiotronic on a 
service contract. 

Sugar did not turn down the proposition immediately, but met 
the people at Audiotronic to discuss what they had in mind. He 
also talked to friends whose business judgement he trusted. One 
Sunday he rang Gulu Lalvani of Binatone and explained the 
proposal. Tm seriously considering it. What do you think?' he 
asked. 

Selling out had not occurred to Sugar before the Audiotronic 
approach; insofar as he thought about the future, he simply 
envisaged a bigger and bigger Amstrad. But the prospect of being 
able to bank a £2 million cheque made him ponder his options, as 
he recalls: 'I wasn't really thinking of cashing in on the thing. But 
I suppose something sparks off in your mind a deal, and you start 
to think maybe the time has come for me to secure myself. It could 
have provided a stepping stone where I wouldn't have had to 
worry about my own personal security.' 

The Audiotronic offer forced Sugar to look at Amstrad from a 
new angle. He had been working his guts out for a decade - and 
for what? True, he now had a more than comfortable lifestyle. 
Amstrad's continuing growth - with profits reaching £649,749 on 
sales of £4,469,396 in the year to June 1978 - ensured that. But 
still he did not have a pile of money in the bank. The Amstrad 
chairman talked through his future with Nick Lightowler at 
Comet: 'Why am I doing this? If I die now, my wife's got nothing,' 
he said. 

The breakthrough came when Sugar discussed Audiotronic's 
offer with Michael Hollingbery, Comet's chairman. He explained 
that he was interested in selling some of Amstrad, but was not 



79 



DAVID THOMAS 

sure whether Audiotronic were the right partners. Sugar's caution 
proved to be amply justified as a succession of disasters struck 
Audiotronic over the following two years: first the Lasky family 
quit the company, then it declared a heavy loss, prompting a sale 
of the Laskys stores to Ladbroke, and finally Rose departed 
abruptly from the scene. 

Sugar told Hollingbery that he was much keener on a partner- 
ship with Comet: T wouldn't mind a deal with you, if you offered 
me the same terms.' 

The Comet chairman was wiser in the ways of the financial 
world than Sugar in 1979. Hollingbery was already a millionaire, 
thanks to the flotation of Comet in 1972, and was to become one 
of Britain's seriously rich businessmen as he sold more of the 
company's stock over the years, culminating in the takeover of 
Comet by Woolworths in 1984. 

With the experience of Comet's own share issue behind him, 
Hollingbery thought there was a better alternative for Sugar. 'Why 
don't you go public?' he asked. 

'Go what?' Sugar replied in his best East End manner, but he 
soon grasped the potential when Hollingbery explained how it was 
possible to tap the stock market for cash. Then Hollingbery went 
one better and introduced Sugar to Tim Holland-Bosworth at 
Kleinwort Benson, the merchant bank which had handled Comet's 
share issue. Thus was Sugar launched on the road to Amstrad's 
flotation. 



A snapshot of Amstrad in 1979, as Sugar prepared to go public, 
would have showed how far the company had moved since 1970 
when he sold his first amplifier. It would also have highlighted the 
characteristic which more than any other marks out Sugar - his 
ability to change both product lines and production sources 
rapidly when fashions change among the thousands of individuals 
who make up his ultimate customers. 

Sugar had introduced more and more products into Amstrad's 
line-up as the 1970s wore on. 'Every year, he produced something 
better and then usually upgraded it, but still aimed it at the same 



80 



ALAN sugar: the amstrad story 

sort of customer. It was nothing for Alan to produce a radio that 
was an absolute winner for a season, but for it then to disappear 
from his offering as fashions changed,' recalls David Johnson, then 
managing director of Rumbelows. 

An important addition to Amstrad's product line had occurred 
in 1975 when it began to import car radios, cassette players and 
speakers from the Far East. Sugar entered the market just as 
demand for in-car music took off. Sales of radio-cassettes for cars 
rose from 276,000 in 1976 to 418,000 in 1977, to 588,000 in 1978, 
and touched a million in 1979. 

Sugar did not just ride the crest of this wave but carried off the 
Amstrad trick of helping to transform the market. He foresaw an 
explosion for low-priced car music systems, an explosion which 
would be ignited by the move towards do-it-yourself installation 
by motorists. Until then, a garage or fitting specialist had charged 
as much as £150 just to install a music system in a car. Sugar 
realized that he could both undercut the competition and create a 
mass market by bypassing the need for fitting charges. Along with 
companies such as Binatone and Sharp, Amstrad was among the 
first wave to market D.LY. packages of radio-cassettes, loudspeak- 
ers and parts for fitting. He also had the sense to sell these 
packages through his normal outlets, like Comet, rather than 
through specialist motoring stores. 

Amstrad had to contend with a setback to its in-car business 
when it was forced by Pioneer - one of the leading Japanese 
companies in the sector - to destroy a batch of speakers which it 
had bought from a factory in the Far East without realizing they 
were directly based on a Pioneer design. This was an incident 
which coloured Sugar's behaviour when he entered the computer 
business: he was extremely careful to ensure that his designs did 
not infringe other people's patents. 

By the end of the 1970s, Amstrad was claiming about 30 per 
cent of the U.K. in-car music market by volume - its market share 
would have been perhaps half that by value, because it concen- 
trated on the inexpensive end. Sugar's ability to spot, transform 
and dominate this business underpinned Amstrad's growth in the 
run-up to flotation. By the time the City heard about Amstrad, 



81 



DAVID THOMAS 

Sugar was doing more business from in-car entertainment systems 
than from his traditional audio sales, representing yet another 
metamorphosis in the company's many different lives. 

In 1977, Sugar turned his mind to selling portable small-screen 
television sets made in South Korea, but the venture was thwarted 
initially by one of the periodic bouts of trade friction between 
Korea and the West. Not to be put off, he returned to the idea the 
following year, also adding clock radios to his new line in portable 
TVs. 

A breakdown of the £4,796,000 of sales notched up by Amstrad 
in the second half of 1979 gives some idea of the range of products 
Sugar was then selling: in-car entertainment systems represented 
30 per cent, the biggest slice, followed by amplifiers, receivers, 
tuners and turntables (19 per cent), clock radios (12 per cent), 
speakers (12 per cent), the new Tower Systems which by then had 
not fully taken off (8 per cent), televisions (7 per cent), cassette 
decks (6 per cent), leaving an assorted mixture of goods to account 
for the final 6 per cent. Not content with that spread, Sugar was 
toying with the idea of making a radio receiver for boats in what 
would have been a concerted Amstrad attack on the high-cost 
marine electronics suppliers - one of the many ventures con- 
sidered, but then dropped in Amstrad's offices down the years as 
Sugar restlessly sought new openings. 

Perhaps even more impressive was Amstrad's range of suppliers, 
which by the second half of 1979 fell into three broad groups. 
First, 49 per cent of its goods came from eight suppliers in Japan, 
Hong Kong, South Korea and Taiwan; factories in the Far East 
were making car music systems, clock radios and televisions for 
the company. Second, 39 per cent of Amstrad's products were put 
together by six U.K. sub-contractors, with L. & N. the largest; 
amplifiers, receivers, tuners, turntables and the Tower Systems 
were among the products assembled in this way. The remaining 
12 per cent of Amstrad's turnover came from speakers which it 
assembled in a factory in Southend. Sugar had bought the 
Southend lease in 1978, thus launching a long Amstrad association 
with Southend and the neighbouring town of Shoeburyness. 

Such a wide spread of suppliers meant that Amstrad was not 



82- 



ALAN sugar: the amstrad story 

over-dependent on any one. Sugar had no long-term contracts 
with any supplier or sub-contractor and was confident of being 
able to find many alternative sources if he fell out with an existing 

supplier. 

In most other respects, however, Sugar's advisers found 
Amstrad to be an admirably simple company as they prepared it 
for flotation. Financial arrangements had remained essentially 
unchanged from its earliest days. Strong growth coupled with the 
lack of a need for substantial capital investment allowed Sugar to 
finance company activities out of cash flow. An overdraft to tide 
Amstrad over the gap between delivering goods to its customers 
and receipt of their payments was the only external financial help 
he needed: for example, he would be paid by Comet in December 
for Christmas stock delivered in July or August. In 1974, Sugar 
had agreed overdraft facilities with Lloyds Bank, giving them a 
claim on Amstrad's properties in the event of the company 
running into financial problems - arrangements which stayed in 
place for the rest of the 1970s. Lloyds extended these credit 
facilities in early 1980 just before the flotation, allowing Amstrad 
a maximum overdraft of £2 million (with interest at 1 >/2 per cent 
above base rate), as well as a £1 million facility to cover letters 

of credit. 

Amstrad had had to change in other respects as it grew from 
being a one-man band to a medium-sized business. The company's 
chronic tendency to outgrow its premises forced another move on 
Sugar in 1977, when he bought premises in Garman Road, 
Tottenham for just over £300,000. This time, the move was 
motivated mainly by Amstrad's desperate need for more storage 
space: Garman Road was little more than a large warehouse 
fronted by a small, two-storey set of offices. 

Sugar also had to take on more managers as Amstrad increased 
in size. Like many a small businessman, his first inclination had 
been to surround himself with family and friends. Back in 1968, 
the documents establishing the company listed three directors: 
Alan Sugar, Ann Sugar and Sugar's brother-in-law Harold, a 
cheerful fellow without any of Sugar's drive who severed his links 
with the company just two months later. Ann Sugar remained one 



83 



DAVID THOMAS 

of the two Amstrad directors throughout the 1970s, although this 
was essentially for tax reasons, since the arrangement allowed her 
to be paid a salary and to be furnished with a Jaguar car. She did 
not play an active role in the company. 

One of Alan Sugar's first employees had been his father, 
Nathan. Sugar saw his own business as a way of freeing Nathan 
from the need to work in the East End garment trade: 

He worked for me because I wanted to stop him working for the 
old sweatshops. He was earning something like £20 a week in 
the garment business, and the way I was doing business, twenty 
quid a week was nothing. So he might as well come and work 
for me and answer my telephones and wrap a few parcels up. 
He was much happier doing that than being under the constant 
threat of having no work in the sweatshops. 

Nathan never quite adjusted to his son's success; he acted as 
though it were built on sand and might come crashing down at 
any moment. When Amstrad was in Ridley Road, he would 
wander the nearby market picking up pieces of string and tying 
them together for Amstrad's parcels. Nathan would sit in the office 
steaming off unfranked stamps to put them on envelopes that 
might contain invoices going out to customers like Comet for tens 
of thousands of pounds. Alan Sugar would battle with his father - 
'Dad, do me a favour - we can afford stamps' - but had no hope 
of changing him. Many a time, Nathan interrupted a meeting his 
son was holding to show him a bargain he had picked up in the 
market: 'Look Alan, look at these lovely oranges - I got you them 
cheap.' 

Derek, Alan Sugar's brother, also started working for Amstrad 
in 1973, supervising the stock in Ridley Road. The plan was that 
if he stuck it for a year, he would move to a more senior position 
and get 5 per cent of the business. However Derek lasted only a 
month with Amstrad before returning to taxi-driving, although at 
the end of the 1980s he helped his brother by managing some 
properties into which Alan had sunk some of his private wealth. 

But Sugar could not rely on his family to staff the whole of 



84 



ALAN sugar: the amstrad story 

Arnstrad and one of the first people he hired was Stan Randall, a 
friend of George Chinchen. Randall helped to shape the company's 
later history by opening AmstracTs first overseas subsidiary in 
Hong Kong in 1981. An open, friendly man, he is a workaholic 
who lives for Amstrad: in Hong Kong in the late 1980s, his 
constant companion was his portable phone in case Alan Sugar, 
or anyone else from Amstrad headquarters, was after him. 

Randall had a small electronics assembly plant in Suffolk when 
he was introduced to Sugar by Chinchen. He was one of the first 
sub-contractors to work on Amstrad's earliest line of audio equip- 
ment launched from Sugar's base in St John Street, Islington. 
When Sugar decided that Amstrad would have to start manufac- 
turing, he turned naturally to Randall. Sugar met Randall and his 
then partner over dinner and offered to take them over, which 
would have given Amstrad ready-made manufacturing capacity. 
Randall's partner declined, but Randall had been sufficiently 
impressed by the young East Ender to decide to throw in his lot 
with Sugar. The 32-year-old Randall started working for Sugar in 
1971 on the day that Amstrad moved from St John Street to Great 
Sutton Street. By then Amstrad had about a dozen employees, 
most of them recently recruited to run its first manufacturing 
operation. 

Stan Randall's first job was to oversee the new manufacturing 
line in Great Sutton Street, but he quickly established himself as 
the company's chief component buyer - chief, that is, after Sugar, 
who was chief everything at Amstrad. Randall's previous experi- 
ence of buying electronic components from the Far East came in 
particularly useful when Amstrad first began to buy large volumes 
of parts from that region. 

Another early Amstrad employee who was to play a key role in 
the company's future was Bob Watkins, a large, engaging man 
from Essex with an accent as broad as Alan Sugar's own and a 
habit of speaking his mind second only to that of the Amstrad 
chairman. In 1975, Watkins was becoming increasingly disil- 
lusioned with his work as a draughtsman for the Ministry of 
Defence. 'I'd been very pleased to get work for the M.O.D. 
because they had such a strong technical reputation. But I couldn't 



85 



DAVID THOMAS 

believe what a bunch of wasters they were. They didn't use their 
brains. Everything I was doing was being repeated and duplicated 
by other departments. Everything took longer than it should have/ 
he recalls. 

Watkins reacted to the Ministry of Defence in much the same 
way as Sugar had to the Department of Education and Science, so 
he looked around for another job and spotted an advert from 
Amstrad, a name which meant nothing to him. T did a bit of 
research and found out that Amstrad was at the rubbish end of 
the market, but I joined just the same,' Watkins says. 

When Watkins started with Amstrad in 1976, there were just 
four people in its technical department: a chief engineer, an 
electrical engineer, himself as draughtsman and an assistant. 
Watkins loathed the place at first, not only because he was earning 
less than in his previous job but also because of the behaviour of 
his new boss. Alan Sugar kept wandering into Watkins's office, 
looking over his shoulder and telling him what to do. This did not 
please Watkins, who reckoned he knew his job. T don't like this 
guy. This is a wind-up,' he thought. 

So he went to Sugar and said he was quitting, but Sugar dissuaded 
him by awarding him an immediate pay rise and suggesting that he 
give the place another couple of months. Watkins agreed, but when 
Sugar started hassling him once more, he threatened to leave again. 
Yet another pay rise kept him at his drawing board. 

This rigmarole might have carried on indefinitely had not 
something clicked within Watkins: 'After a few months, I suddenly 
realized this guy was a bit clever. Some of his ideas were daft, 
O.K., but some of them were better than I could have thought of 
and I reckoned I was pretty good. My experience of bosses up till 
then was that they didn't know what they were talking about. But 
here was a boss who was different.' 

Watkins was less inclined to take his initiation into Amstrad 
personally when he noticed that he was not alone in receiving the 
Sugar treatment. Sugar was involved in everything that went on 
in the company, from the accounts, through the warehousing and 
the manufacturing to the work of the electrical engineers. And, as 



•86 



ALAN sugar: the amstrad story 

far as Watkins could tell, most of what Sugar had to say in these 
very different areas made sense. 

Once he had become used to his unusual boss, Watkins found 
Amstrad's way of operating a great relief after the bureaucracy of 
the Ministry of Defence. If anyone within Amstrad showed a 
talent for something, they tended to do it, irrespective of their job 
title. After a while, he discovered he had a talent he had never 
suspected: he was rather good at hunting out cheaper sources of 
supply for the various parts such as metal work and plastics that 
Amstrad was buying from U.K. manufacturers. Watkins is a street 
fighter like Sugar, a man who knows where to find the cheapest 
sources of supply, understands about tough negotiating and can 
almost instantaneously calculate the likely cost of a new product. 

Recognizing Watkins's skill, Sugar began to take him on his 
expeditions to the Far East where he bought components and 
negotiated deals with his sub-contractors. Gradually Watkins took 
over from Randall as Sugar's right-hand man in this key aspect of 
Amstrad's operations. 

Watkins's main responsibility at Amstrad has always been for 
new products. In the 1970s, Sugar invariably came up with all the 
new product ideas and would normally have strong views about 
how they should look. He would use Watkins as a sounding board 
to discuss these ideas and the two men found that they thought in 
similar ways. Both wanted to design simple low-cost items. And 
Sugar impressed on Watkins his philosophy of new products: 
fc Once we've come up with a product, we have to assume it's old 
and will have stopped selling next year. It may turn out to have a 
longer life, but that doesn't matter. We have to assume it's already 
old-fashioned.' 

This was a lesson which Watkins did not forget even after he 
joined Amstrad's board as the company's technical director in the 
1980s: 'Once an Amstrad product has come out, I have absolutely 
no interest in it,' he says. 

As each wave of Amstrad's new products tended to sell in ever 
larger quantities, Alan Sugar realized that he needed help on the 
marketing side. In 1977 he began to talk about finding some bright 
lad w r ho could give him a hand. Malcolm Miller was such a lad. 



87 



DAVID THOMAS 



Miller had just finished a degree in business studies at the Central 
London Polytechnic, which had included a year working on 
marketing for Birds Eye - a flagship brand in Unilever, the Anglo- 
Dutch group which is widely considered one of the world's most 
professional marketers. Miller heard through a friend of a friend 
that Sugar wanted to appoint a marketing professional, so he rang 
up and was interviewed by the Amstrad boss. 

Sugar would not offer the salary Miller thought he deserved, so 
the young graduate took up Birds Eye's offer of a full-time job. He 
spent the next year and a half learning about product management 
and marketing on Unilever's training scheme, which has been 
compared in its intensity with an M.B.A. (Master of Business 
Administration) in marketing. 

After about eighteen months with Birds Eye, Miller decided 
that the clever thing to do would be to apply the marketing skills 
he had learnt at Unilever within a small or medium-sized com- 
pany. He saw a vacancy for a product manager at Amstrad 
advertised in London's Evening Standard newspaper, and this time 
took the job when it was offered. It was 1979, Miller was 24 and 
Amstrad had hired its first marketing executive. 

In the 1980s, when Miller had joined Amstrad's board as its 
sales and marketing director, he w 7 as often privately described by 
the company's critics as Sugar's chief clone. Like a handful of 
other senior Amstrad people, he happens to look like the Amstrad 
chairman - taller and gaunter, but with the same mop of black 
curly hair and the same beard. On occasions, he can also affect 
Sugar's brusque manner. Yet what the clone jibe misses is that 
Miller brought into Amstrad big-company marketing disciplines 
which complemented Sugar's gut feeling for what the consumer 
wants. 

When Sugar wrote to Miller offering him a job, he told him that 
he would have to learn a hell of a lot and no one would have the 
time to teach him. Miller soon found out what his new boss meant. 
In his first few months with Amstrad, the young man had to 
handle the company's advertising, link with its retail customers, 
telex Far Eastern suppliers about shipments of speaker parts, open 
letters of credit, write instruction manuals and design gift boxes 



88 



ALAN sugar: the amstrad story 

for promotional launches of Amstrad's products. More mun- 
danely, Miller also had to type his own letters, since Sugar made 
it clear that a secretary was an unnecessary luxury. It was all 
rather different from Unilever, where everyone had had their 
precisely allotted function. 

Another big difference from Unilever was that no one outside 
the confines of the audio industry had ever heard of Amstrad: 
'When I phoned up an agency or anyone and I said I was from 
Amstrad, they'd say: "Who's Amstrad?'" 

That was galling enough for an ambitious young man like 
Miller, but more worrying still was the complete lack of anything 
approximating to conventional marketing disciplines within 
Amstrad. The company's entire marketing strategy was locked in 
Sugar's fertile mind: 'The first thing I said to Alan was: "Look, we 
don't know who our customers are - what they do, where they 
live, what papers they read,'" Miller recalls. 

An early change suggested by him was to include a question- 
naire with Amstrad's guarantee card, so that the company could 
begin to gather some intelligence on the people who bought its 
equipment and how they had heard about it. The replies which 
flooded in soon indicated at least one thing: that the television 
advertising campaign for the Tower System had been an over- 
whelming success. 

One of Miller's early tasks was to keep an eye on the shooting 
of a television advert for the Tower System. Sugar had decided 
that since he wanted a mass market, he ought to use the 
advertising medium that talked to the great majority of people - 
television. Amstrad was one of the first companies to advertise a 
brown goods product like an audio system in such a direct and 
simple way. 

The commercial which Miller saw being shot was pure Sugar. 
The camera travelled through space, breaking the £500 barrier, 
the £400 barrier, then £300, coming to rest below the £200 figure. 
By the end of the advert, no one could have the slightest doubt as 
to what product was being advertised, where it could be bought 
and how much it cost. 'When we researched that advert, we found 
that consumers really appreciated it. Don't forget we were talking 



89 



DAVID THOMAS 

to C2s and DEs (working-class consumers). ABs (the professional 
classes) had bought their Sony and didn't mind paying a high 
price. C2s and DEs wanted something that looked like hi-fi 
separates, but cost them £179,' Miller explains. 

Simple to the point of being corny, perhaps, but the advert 
worked. Amstrad received feedback from outlets in the Midlands 
and the North that whenever it ran on a Saturday afternoon, a 
surge of customers would come into the shops to buy a Tower 
System an hour later. The power of advertising - especially large 
volumes of simple advertising on television - was a lesson that 
Amstrad would never forget: The correlation between advertising 
and our product offtake is still phenomenal,' Miller says. 

As a growing company, Amstrad also began to need more 
conventional financial controls. In 1977, Sugar asked Nevill Shear- 
man - a partner in Amstrad's auditors, Mordant Latham - to find 
him a financial controller. The advertisement which Shearman 
placed in the Daily Telegraph attracted Jim Rice, then working with 
a Swiss cosmetics firm. Rice did not have an accountancy qualifi- 
cation, but he had worked in the financial departments of a range 
of large companies including A.E.I., the engineering group, and 
Philips, the electronics giant. 

Rice's first task in 1977 was to introduce elementary accounting 
systems into Amstrad and to help computerize its administrative 
operations. As is the way with Amstrad, Rice's job soon spread. 
In the late 1970s and early 1980s, he assumed responsibility for 
most aspects of Amstrad which were not front-line selling activi- 
ties, like warehousing and manufacturing. When Amstrad applied 
to join the stock market, it needed a more conventional board 
structure, so Rice began to wear yet another hat - that of company 
secretary and main contact point with the City. 

When Rice joined Amstrad, the company's headquarters were 
still in the centre of Ridley Road market and, like many newcomers 
to the firm, he took a good six months to adjust to his new 
surroundings. Tt was a cultural shock after working for someone 
like Philips. Philips is very well structured - very slow and 
cumbersome, but very well structured. In Philips, everybody knew 
their status. If you had two windows in your office, you'd made it. 



•90- 



ALAN sugar: the amstrad story 

If you had a window at all in Ridley Road, you were lucky,' recalls 
Rice. 



Alan Sugar was beginning to build up a team, many of whom 
were to play important parts in Amstrad's success in the 1980s. 
Yet he himself continued to provide the driving force in all aspects 
of company operations. The memo pad on his desk in Tottenham 
in 1979 proclaimed: 'From the Boss'. Nowhere was this more true 
than in relations with Amstrad's main customers, the large chains; 
Sugar reserved for himself contact on all but the most trivial 
matters with retailers like Comet and Rumbelows. 

David Johnson at Rumbelows says that it was nothing for Sugar 
to visit his offices two or three times a month to sort out problems. 
Differing from the heads of other Rumbelows suppliers in never 
standing on ceremony, the Amstrad chairman would deal with 
whoever was most appropriate, unlike many rival suppliers who 
felt their dignity deserved personal contact with managing director 
Johnson. 'Alan didn't mind who he talked to. If there was a 23- 
year-old in Rumbelows who was important for a particular line, 
you'd find Alan stuck in some cubby-hole talking to the lad about 
how the sales were going, what improvements could be made and 
so on,' Johnson remembers. 

One evening just before Christmas, after most of the staff had 
gone home, Johnson came down into the lobby to find Sugar 
sitting beside the reception desk: 'Alan, what on earth are you 
doing here?' Johnson asked. 

T've come for the cheque,' Sugar replied, explaining that 
Rumbelows still owed him for the Christmas season. He saw no 
reason why the money should be sitting in their bank account over 
Christmas earning interest, when it could be in Amstrad's, and 
was prepared to wait for as long as it took Rumbelows's finance 
director to sign the cheque. 

Sugar addressed the early preparations for the flotation with 
that same single-mindedness. When a team from Touche Ross, 
who became Amstrad's auditors in the run-up to flotation, first 
visited the Tottenham headquarters, Sugar introduced Malcolm 



91 • 



DAVID THOMAS 

Miller to them: 'Malcolm's our product manager on the marketing 
side. He's worked for Unilever and he's a graduate. He came to 
me about two years ago looking for a job, but I said to him, "No, 
you could be more valuable to Amstrad if you worked for a big 
company like Unilever first. Come back after you've had a couple 
of years experience and then join us.'" 

Miller had to suppress a smile at this new version of how he had 
joined Amstrad - a story which quickly became the authorized 
version for anyone connected with the company's share sale. Sugar 
also needed publicity material for the flotation process which 
would show where Amstrad's machines were made. So he travelled 
down to Ron Nixon's factory in Rochester and, for the photogra- 
pher's benefit, temporarily replaced the L. & N. sign on the factory 
with an Amstrad one. Nixon asked Sugar whether the effort in 
preparing for the flotation was worth it. 'Once we've gone public, 
I'll have a million quid of me own and so will the missus. Then 
I'll be able to gamble with someone else's money,' Sugar replied, 
breaking into one of his grins. 

Yet Sugar had no intention of putting his feet up after Amstrad 
went public, even though it guaranteed his own security and that 
of his family. As he told the Evening Standard when asked how the 
money would change him: 'I'm not going to book into a suite at 
the Hilton and take things easy for the rest of my life. Don't forget 
I've still got the business to run.' 



•92 



In the Public Eye 



Launching into a recession is not the best start a consumer 
electronics concern could wish for as a public company. Con- 
sumers who feel the cold wind of economic problems blowing 
around them tend to shelve plans to buy the kind of equipment 
Amstrad sold back in 1980 - hi-fi units, car radios, small-screen 
televisions and so on. Yet recession is precisely what Amstrad had 
to contend with in its first two years as a public company. 

Sir Geoffrey Howe unveiled £3 billion of public spending cuts in 
his second budget as Chancellor of the Exchequer on 26 March 
1980, a month before Amstrad's late April flotation. In August, 
unemployment passed the 2,000,000 mark for the first time since 
the 1930s and the gloom continued almost unabated for two years: 
in the first half of 1982, the Confederation of British Industry 
called for an emergency £3 billion package of government spending 
to help industry and released one of its most pessimistic surveys of 
business confidence. 

Alan Sugar was prudent enough to bend his plans to the 
prevailing economic climate. In the summer of 1980 he cut stocks 
and postponed the launch of a new hi-fi system. Yet more 
symptomatic of Amstrad's fortunes at the time was the announce- 
ment in October 1980 that the firm had beaten the profit forecast 
made by Sugar before flotation: pre-tax profits for the financial 
year 1979-80 leapt 50 per cent to £1.36 million on sales up 56 per 
cent at £8.76 million. 



93 



DAVID THOMAS 

Growth on that scale was to be the pattern for the next few 
years. The problem Amstrad faced was not one of demand, but 
rather of increasing its sources of supply. Sugar seemed to be 
almost semi-permanently engaged in a search for larger warehous- 
ing and factory capacity. By October 1982, he was feeling suf- 
ficiently confident to observe that the recession had passed the 
company by. He even wondered why the C.B.I, never consulted 
Amstrad before producing its gloomy accounts of the economy, 
furnishing the answer to his own question by observing that 
perhaps Amstrad was, after all, still too small. 

Knowing with hindsight what was in store for Amstrad later in 
the decade, it is tempting to dismiss the period between 1980 and 
1983 as an irrelevant interlude. But that would be wrong: for these 
years, when Amstrad was fully exposed to the public view for the 
first time, revealed some of the company's main characteristics. 
Sugar demonstrated immense agility in the way he moved into 
and out of markets; he laid important foundations for Amstrad's 
future growth; and he began to articulate a distinctive Amstrad 
business philosophy. Moreover, relations with the City were set on 
the fraught path they have followed ever since. 

Sugar's marketing agility was most clearly demonstrated during 
the early 1980s by the way he coped with Citizen's Band radio. 
Britain was one of the last countries in Europe to legalize the craze 
among radio hams for endlessly chatting to each other over the 
air-waves - a craze which had become a national obsession in the 
United States. The British Government had been dithering for 
more than two years over what form of CB system to allow. 
Amateur radio enthusiasts had lost patience: a million CB radios 
were thought to be illegally occupying air-wave space. 

Interest in the potential CB market was intense by the end of 
1981, as it became clear that the Government was finally lumber- 
ing up to a decision. About twenty companies including Amstrad 
were poised to enter the market, mainly with CB sets imported 
from the Far East. Optimistic projections abounded: one pundit 
predicted sales of a million sets in the first year, but added that 
the market would not be saturated for a further five years, when 
six million CB radios would be sold annually. 



94 



ALAN sugar: the amstrad story 

Alan Sugar had been keeping half an eye on CB for a long time; 
before Amstrad's flotation, he had declared his interest in jumping 
on the band-wagon as soon as the hobby was legalized. But he 
also made plain his scepticism of the rosiest projections of CB's 
potential. As he wrote in his chairman's statement accompanying 
Amstrad's 1980-81 report and accounts: 

It is a market which in other countries has caused several 
failures, due to the bad quality of products and a poor under- 
standing of the demand, and its recent legalization may incite 
some over-enthusiasm on behalf of some suppliers in respect of 
market size and potential. It is therefore our intention to enter 
the market . . . with a degree of caution in our commitment. We 
will try to adopt a sensible medium whereby we can react either 
way to its future. 

The statement was dated 6 October 1981, five days after the 
Government had announced plans to legalize CB. 

Sugar suspected that CB would be a one-week wonder. He knew 
there was a reservoir of pent-up demand from frustrated would-be 
radio hams, but he also reckoned that there was a distinct limit to 
the number of Britons who wanted to spend their leisure hours 
swapping call signs over crackling air-waves. Armed with this 
intuitive feel for the market, he concluded that the way to make 
money out of CB was to be there early, in large volumes, but also 
to be ready to quit the minute the market looked like drying up. 

Amstrad was one of the first companies to secure large orders, 
running into tens of thousands of sets, from chain stores like Argos, 
Currys and Rumbelows, who were busy stocking up for a 
Christmas CB boom following the October legalization. Sugar was 
also the first big player in the market to spot the end of the craze 
early in 1982; as soon as he realized the CB bubble had burst, he 
moved quickly to ensure that all the sets which the retailers had 
ordered from him were safely booked into their warehouses. By 
contrast, most of Amstrad's main competitors were stuck with 
large stocks and heavy losses. 

Gulu Lalvani, Sugar's old friend from Binatone - which also 



95 



DAVID THOMAS 

went for CB in a big way - says: 'Alan smelt, I mean smelt, that 
the market was on the point of going down. And his reaction was 
incredibly sharp.' 

A couple of months later, when everyone had realized that CB 
was fizzling out, Lalvani asked Rumbelows to take some of his 
remaining stock at discounted prices. 'I was told by Rumbelows: 
"We can't take another set, because our warehouse is bursting 
with Amstrad CB radios." That was Alan's sharpness,' Lalvani 
recalls, with a rueful smile. 

Meanwhile, Sugar was shifting his Tower Systems and a host of 
new audio products in ever-increasing numbers. In retrospect, it 
is easy to see that Amstrad's product line was ideally placed to 
benefit - not suffer - from the recession. Consumers feeling the 
pinch were as likely to opt for value-for-money as to abandon their 
purchasing plans altogether. They chose a music system a little 
less expensive than they had originally planned. In February 1982, 
the Daily Telegraph explained Sugar's ability to buck the trend of 
the recession as 'deliciously simple': Amstrad was producing 'an 
audio entertainment system that looks like it should cost £400, to 
sell at between £150 and £200'. 

The financial pressure on the working- and lower-middle-class 
customers who tended to buy Amstrad's equipment was eased a 
little in 1982 when the Government abolished restrictions on hire 
purchase. The company was waiting to take their business and 
Sugar made sure they knew about Amstrad by pouring ever-larger 
sums of money into advertising. Some £3 million went on television 
advertising and full-page ads in all the popular daily newspapers 
in the Christmas season of 1982, for example, as compared with 
Amstrad's total annual advertising budget of about £100,000 in 
the late 1970s. Fuelled in part by this television exposure, Sugar 
was able to claim a third of the 850,000 hi-fi units sold in Britain 
in 1982 - much more than any competitor, including the Japanese. 

Many business people would have been stretched simply to keep 
up with such a boom in their core business, but Sugar was careful 
during these early years in the public eye to sow seeds designed to 
maintain Amstrad's rapid growth. One important area was his 



96 



ALAN sugar: the amstrad story 

continuing struggle to diversify the company's spread of cus- 
tomers, thereby reducing its over-reliance on Comet. The obvious 
mass market appeal of the Tower System allowed Amstrad to 
penetrate the mail order business by forging links with mail order 
leaders such as Grattan. Sugar consolidated his one substantial 
export market by opening a French subsidiary in 1982. Sales to 
high street chains like Rumbelows, Woolworths and Currys also 
flourished, to the point where by 1982 Amstrad's equipment was 
being sold through 3,000 retail outlets. 

Sugar realized that his manufacturing arrangements would have 
to evolve if he was to sustain Amstrad's growth. He could no 
longer rely on a clutch of small British sub-contractors, like 
L. & N. in Rochester, whose output could not always be guaran- 
teed because of their perennial struggle to recruit enough skilled 
labour. Sugar broke the news to Ron Nixon at L. & N. that 
Amstrad would no longer be putting work his way. Nixon then 
experienced - for the first and last time - Sugar's tendency on 
occasion to stick so rigidly to the letter of an agreement that the 
person on the receiving end perceives it as ruthlessness: a tactic 
which he is particularly prone to employ when he has no further 
business with someone. 

While assembling an entire piece of equipment for a customer, 
a sub-contractor like L. & N. inevitably has to contend with 
mismatches in the component parts supplied by the customer. 
L. & N. needed more than 1,000 sets of components to make 1,000 
of Amstrad's products, because individual components were lost 
or damaged during the manufacturing process. When Sugar 
wound up Amstrad's dealings with Nixon, he charged L. & N. for 
all the components that could not be accounted for down the 
years. 'It's normally accepted with everyone else I've worked with 
that they give you a bundle of stuff now and again to offset 
component losses,' Nixon says. 

Insisting on what was undoubtedly Amstrad's legal rights, 
Sugar refused to follow suit. 'It chewed away about half the money 
that they owed us in the end. And it was on that that we came to 
blows virtually,' recalls Nixon, before adding, 'Mind you, that was 
the only example where I considered him to be unfair.' 



97 



DAVID THOMAS 



In place of L. & N. and other similar sub-contractors, Sugar 
relied on two centres of manufacturing. Amstrad's first and most 
important manufacturing base was in the Far East. In 1981 Sugar 
strengthened his links with the region by opening, a Hong Kong 
office, the company's first overseas subsidiary. Secondly, he built 
up his U.K. capacity, initially in 1981 by leasing a 40,000-square- 
foot factory in Southend near the one he had leased in 1978. When 
Amstrad outgrew that too, he paid £2.5 million in 1982 for a 
133,000-square-foot plant at Shoeburyness, on the Essex coast. 

Stan Randall, who had been one of Amstrad's first employees 
when he joined the company a decade earlier, was going through 
a separation from his first wife in early 1981 which he describes as 
'ugly'. A strong family man, Sugar did what he could to help 
Randall over a difficult patch and approached him one day in 
August with a proposal: 'How would you like to go to the Far East 
for a couple of years to help you get over it? I think I can get you 
out there fairly soon.' 

Sugar explained to Randall that he had been thinking of setting 
up a Hong Kong office, which would allow Amstrad to cut out its 
Japanese agent. In future, it would deal directly with all its sub- 
contractors and suppliers, which would not only save on the 
agent's mark-up but also give Amstrad greater control over 
suppliers. Sugar planned to negotiate the main contracts with sub- 
contractors and suppliers during his frequent visits to the Far 
East, but Amstrad's Hong Kong office would be in a better 
position to keep in day-to-day touch with its Far Eastern suppliers, 
not least because it was in the same time zone. In addition, 
Randall in Hong Kong would be able to oversee the shipping of 
Amstrad's goods to Britain. 

So keen was Randall to accept Sugar's offer that by October 
1981 he had started work as the first employee of Amstrad 
International (Hong Kong) Limited. Amstrad occupied a small 
corner of a textile company run by Simon Mak, a Hong Kong 
Chinese businessman. It was thanks to Gerry Eriera, Sugar's 
Chigwell friend, and Ann Sugar that Alan had met Mak, the 
agent through whom Eriera imported materials for his ladies' 
fashion business in which Ann had taken an interest. Sugar and 



•98 



ALAN sugar: the amstrad story 

Mak hit it off so well that he asked Mak's help when setting 
up the Hong Kong office. Stan Randall launched Amstrad's 
Hong Kong operation surrounded by Chinese haggling over 
cloths: 'I was given a desk and a chair and a very tiny office, 
with a telephone line, and I was expected to get on with it,' he 
recalls. 

Randall hired a secretary recommended to him by Simon Mak, 
but she lasted only two weeks because she had worked in Britain 
and therefore was used to a 9-5 routine. As his new secretary he 
employed Callen So, a 19-year-old Hong Kong Chinese woman 
who quickly proved to have the essential Amstrad qualities: an 
appetite for extremely hard work and the ability to turn her 
attention to almost any problem. The Hong Kong office's 
responsibilities expanded steadily. Amstrad began to arrange 
some of its letters of credit in the Far East, leaving Randall to 
discuss terms with bankers there. In addition he hired design 
and engineering staff who often acted as intermediaries between 
Hong Kong and Amstrad U.K. After Amstrad France was 
launched in 1982, the Hong Kong office also started to arrange 
shipments direct to France, saving on the time and money 
involved in shipping via Britain. In the early years, Callen So 
often worked in Amstrad's little offices till 9 or 10 in the evening. 
Her mother began to scold her, saying these hours were not good 
for a young woman and asking her to quit Amstrad. But she had 
caught the Amstrad bug and Sugar rewarded her later in the 
decade first by making her Amstrad's marketing director in the 
Far East, and then by bringing her to Amstrad headquarters as 
his special assistant. 

As far as manufacturing was concerned, Amstrad Hong Kong's 
relations with Amstrad U.K. quickly settled into the groove they 
were to occupy for much of the 1980s. Most Amstrad products 
w r ere either made by sub-contractors in the Far East or were 
simply badged goods imported from Far Eastern OEM manufac- 
turers. The Shoeburyness plant acted as a large warehousing 
centre for the distribution of Amstrad's goods arriving from the 
Far East. It was also used for putting the finishing touches to 
goods made there and for assembling into finished form products 



99 



DAVID THOMAS 



whose bulk made them uneconomic to import. Amstrad had 
originally acquired factory space in Southend back in the 1970s to 
assemble bulky products like loud-speakers. The large-screen 
televisions which the company started selling in 1983 were one of 
the first products to be assembled in Shoeburyness. 

Alan Sugar also spent the early years of the 1980s restlessly 
thinking about new products. He constantly unveiled new models 
for Amstrad's audio line-up, for instance adding a remote control 
to the Tower System, a feature which in those days tended to be 
found only in hi-fi systems in the £1,000 bracket. Yet he knew that 
Amstrad could not realistically expect to take an ever greater share 
of the U.K. audio market. Moreover, by the end of 1982 he had 
already dropped two of the main product lines important to 
Amstrad at the time of its flotation: in-car music systems and a 
range of small electronic goods such as portable and clock radios. 

It was imperative for Amstrad to build up product areas outside 
audio. Fundamental to Sugar's business philosophy is the maxim 
of never becoming too dependent on any particular product. The 
Tower System was a winner for the moment, but he knew that its 
very success would encourage others to imitate it. The resulting 
pressure on Amstrad's market share would inevitably force him to 
cut prices and margins. 

Sugar's first two entirely new product offerings in these years 
proved a dead end, for the time being at least. In 1982, Amstrad 
began to import badged 14" and 20" colour televisions from the 
Far East. They sold well and the 14" version was voted joint best 
buy by Which? magazine in February 1983. He decided that 
Amstrad should go one better and assemble its own 22" colour 
TVs. When it opened the Shoeburyness plant in 1983 a line was 
installed which could produce 900 TVs a day. Sugar also decided 
to launch into video recorders in 1983, then a relatively new 
product. He began by importing complete video recorders from 
Japan, but laid plans to assemble these in Shoeburyness too from 
mid- 1984. Taken together, he realized, colour TVs and video 
recorders accounted for a very large slice of all consumer electron- 
ics sales and were also likely to furnish much of the growth in the 
market in the immediate future. In October 1983, he talked of 



100 



ALAN sugar: the amstrad story 

winning 10 per cent of the £1 billion annual sales of video recorders 
and 22" TVs. 

But this was not to be. Very few European companies were 
managing to make money from colour televisions in the face of the 
onslaught from Japanese companies, which regarded output of a 
million sets a year as the minimum necessary to guarantee 
adequate economies of scale. Towards the end of 1983, the 
Shoeburyness TV production line also suffered technical problems 
which occupied much of Bob Watkins's time. With video 
recorders, Amstrad felt the backlash of an early round of trade 
friction between the European Commission and Japan: the EC 
included the kits of video recorder parts, which Amstrad wanted 
to import, in quota arrangements designed to protect domestic 
European video recorder producers against the wave of Japanese 
imports. Sugar bowed to the inevitable, announcing his decision 
to run down the company's activities in large-screen colour TVs 
and video recorders in October 1984. 

Predicting a market glut in both colour TVs and video 
recorders, he said in his chairman's statement accompanying the 
1983-84 results that Amstrad 'would be more comfortable as an 
observer rather than a participant in what could become a price 
war'. He underlined Amstrad's pragmatic attitude to products, 
adding that his decision might turn out to be temporary: 'We have 
the flexibility to turn off our supplies when we consider the market 
is vulnerable.' 

In his chairman's statement in October 1984, Sugar justified his 
withdrawal from the video recorder and large-screen TV market 
in terms of Amstrad's overwhelming commitment to the bottom 
line: 

I have always tried to emphasize how margin is important to 
us; we shall only continue in sectors where there are good 
margins of profit. It may seem surprising that one year we are 
selling large quantities of VCR and the next we sell none. Quite 
frankly, when there is no margin in a product, we see no reason 
to endanger our cash flow simply to impress the compilers of 
market share statistics. 



101 



DAVID THOMAS 



By the time he made these comments, however, Sugar had 
launched Amstrad into a product line - personal computers - 
which was to open a new chapter in the Amstrad story. 



A curious episode in Amstrad's history - which resulted in two of 
Britain's most eminent barristers arguing over the company's 
conduct before the House of Lords - can also be traced back to 
the early 1980s. As with most complex legal proceedings, the case 
of Amstrad versus the music industry stretched out over several 
years and was not finally resolved until 1988. But the issues it 
raised for the company stemmed from decisions taken before it 
launched into computers. 

Amstrad fell foul of the music industry's determination to stamp 
on individuals making copies of music cassettes in their homes. 
This is technically illegal in Britain because it infringes copyright 
legislation. The music companies claim that their losses through 
home taping run into millions of pounds each year. Unfortunately 
for the music business, however, the law cannot be enforced 
against individuals and in the early 1980s the industry hit upon 
another tactic: to make an example of an audio manufacturer who 
brought out a high-speed tape-to-tape copier, then a relatively 
new high street product. 

The British Phonographic Industry, the music industry's trade 
association, thought it had caught its prey in 1982 when Aiwa, the 
Japanese audio company, launched a high-speed tape deck which 
copied tapes at four times the normal speed. But Aiwa agreed not 
to sell its recorder in Britain when warned that the B.P.I, would 
make it a test case: like many Japanese companies, it was sensitive 
about its public profile. 

Amstrad was made of sterner stuff. The company's interest in 
the new tape-to-tape copying market dated from 1980, when it 
released a double cassette tape recorder which allowed customers 
to copy from one tape to another and record off the radio or 
records - all for less than £100. But the music industry really woke 
up to Amstrad in 1984, when the company issued three new tape- 
to-tape recording machines which could record at twice normal 
speed and sold for £150 to £200. 



102 



ALAN sugar: the amstrad story 

As is its way, Amstrad was not shy about broadcasting the 
merits of its new machines and launched a big television and 
popular newspaper advertising campaign. 'You can even make a 
copy of your favourite cassette,' the ads stressed, just in case 
anyone was dumb enough not to grasp the point of the machines. 

Japanese manufacturers already had recorders on the market 
which could copy at double speed, but Amstrad's machines were 
in a different league as Patrick Isherwood, then the B.P.I.'s legal 
director, explains: 

Amstrad's advertising was totally different to anything that had 
been seen before. They were spending huge amounts of money 
on national TV at prime times - 'Coronation Street', 'News at 
Ten' and so on. The second difference was the size of their 
market. Amstrad's whole business was geared to large turnover 
at low price. Before, relatively speaking, the quantities weren't 
large. The Japanese machines were usually separates designed 
to go into hi-fi systems. Amstrad took it into the Tower Systems 
market. 

In October 1984 the B.P.I, wrote to Alan Sugar and to eleven of 
Amstrad's main high street customers, pointing out that home- 
taping was illegal and that encouraging people to tape at home 
was potentially illegal in itself. Sugar's response was immediate: 
he issued a writ against the B.P.I, on the grounds that the trade 
association had falsely accused Amstrad of behaving illegally. He 
also sought a declaration from the High Court that it was doing 
nothing wrong. 

The two sides did not meet each other throughout the legal 
twists and turns of the next four years. Sugar never needed to 
appear in court. It was a highly professional engagement, with no 
room for compromise between two fundamentally different objec- 
tives. Amstrad wanted the shadow cast over itself and its cus- 
tomers lifted, while the music industry was trying to reduce home 
taping through a clarification of the law. Indeed, so keen was the 
music business to have the law clarified that private worries were 
expressed within the industry about the health of its opponent. At 



103 



DAVID THOMAS 

a meeting of the B.P.I.'s governing council in late 1984, one of the 
assembled music moguls voiced concern about Amstrad's staying 
power: it was only an insignificant little outfit, so how could they 
be sure that it would still be around in a few years' time to 
continue a legal fight against the massed ranks of the world's 
biggest music conglomerates? Shortly afterwards, Amstrad's sales 
went through the roof and Sugar emerged as a star of the business 
scene: by luck the music industry had chosen the perfect company 
to ensure that its anti-copying case received maximum publicity. 

The first round went to the music business when the case 
reached the High Court in June 1985. Mr Justice Whitford was 
unimpressed by a note Amstrad stuck on its twin-cassette decks, 
which read: 'The recording and playback of certain material may 
only be possible by permission. Please refer to the Copyright Act 
1956 and the Performers Protection Acts 1958 and 1972.' Stating 
that the note was 'barely noticeable or unintelligible to ordinary 
people 1 , the judge dismissed Amstrad's request for a declaration 
and found for the music industry on all points. 

The music industry was naturally cock-a-hoop; immediately 
after the judgement the B.P.I.'s Patrick Isherwood said, Tt will be 
difficult for retailers to sell twin-deck cassette machines after this 
decision. . . . They will have to show adequate warning notices 
which will be off-putting to most consumers.' 

Sugar cast a different light on the judgement and accused the 
judge of 'losing his bottle', explaining this unusual aspersion on 
the English judiciary: Tt seemed to me he was being evasive. He 
wanted to force us to take the matter to the Court of Appeal, 
because it is such a fine point.' 

He was not planning a retreat. 'What the judge has done is so 
anti-Amstrad that we have no choice but to appeal. We have no 
intention of stopping selling the decks,' he said at the time. 

At this point the legal process moved with unusual speed. It 
took only four months for the case to reach the Court of Appeal, 
where the three judges distinguished themselves by producing a 
ruling which satisfied neither side. All the findings of the High 
Court were reversed. Lord Justice Lawton ruled against the B.P.I, 
on the grounds that Amstrad was not authorizing people to copy 



104 



ALAN sugar: the amstrad story 

tapes simply by advertising the copying ability of a particular 
machine. This insight - that Amstrad was not itself authorizing 
home- taping - was to prove fundamental to the rest of the case. A 
parallel would be this: a gun manufacturer does not authorize 
people to commit murder simply because his products can be used 
for this purpose. 

Yet still the Court of Appeal would not give Amstrad a 
declaration that it was doing nothing wrong. On the contrary, the 
Court found that Amstrad might be committing the criminal 
offence of inciting people to break copyright law, but that this 
would have to be settled by a criminal prosecution and not by the 
civil action then under way. 

Amstrad had failed to get its declaration, but the B.P.I, had not 
squeezed out of the Court the simple statement of the law it was 
seeking, as Patrick Isherwood recalls: The B.P.I, had not gone 
into this as a way of punishing Amstrad, but in order to get the 
law clarified. So it was a hollow victory for the B.P.I.' 

There the matter might have rested except for the fact that a 
group of music companies led by C.B.S., the world's largest music 
group, had launched a parallel action claiming damages from 
Amstrad for copyright infringement. This had lapsed while 
Amstrad's request for a declaration was wending its way through 
the courts, but now that this action had come to a halt the C.B.S. 
case was revived, eventually reaching the Court of Appeal in 
February 1987. 

The conclusions of the Court of Appeal on this occasion did not 
differ markedly from those of the same Court two years previously. 
By a two-to-one majority, the Court found against the music 
companies, ruling that incitement to infringe copyright is a 
criminal matter not subject to the civil courts. Lord Justice 
Nicholls told the music companies that their remedy lay in 
launching a private prosecution against Amstrad. Sir Denys 
Buckley - the dissenting judge, who argued that the record 
companies should be allowed to sue Amstrad - quoted a precedent 
dating from 1861 when the Emperor of Austria had been allowed 
to sue individuals who had imported fake bank-notes into 
Hungary. 

This was the first and last time that the Emperor of Austria 



105 



DAVID THOMAS 



entered the Amstrad story. It took another fifteen months for the 
C.B.S. case to move from the Court of Appeal to the House of 
Lords, the highest court in Britain, where lawyers were treated to 
the delicious spectacle of two of Britain's foremost QCs - Robert 
(now Lord) Alexander for Amstrad, and Sydney Kentridge for the 
music companies - dissecting ever more abstruse points of law. 
The Law Lords ruled in favour of Amstrad, reaffirming the 
fundamental point: Amstrad was not guilty of a breach of copy- 
right because it was not authorizing the people who bought its 
recorders to copy tapes. Within the hi-fi industry, the ruling 
evoked a sigh of relief - albeit a rather discreet sigh, as leading 
audio journalist Arthur Ord-Hume recalls: 'Alan Sugar was the 
only one with the gumption to take on the music companies. The 
entire industry went into a huddle and toasted him. The ruling 
was particularly helpful to the Japanese, who had been too 
cautious to take on the B.P.L legally.' 

Yet the Law Lords coupled their judgement with a ringing 
denunciation of existing copyright law as unenforceable. Indeed, 
Lord Templeman became quite lyrical on the point: 'Whatever 
the reason for home copying, the beat of "Sergeant Pepper" and 
the soaring sounds of the "Miserere" from unlawful copies are 
more powerful than law-abiding instincts or twinges of conscience. 
A law which is treated with such contempt should be amended or 
repealed.' 

Once more, both sides could claim a victory of sorts. Nothing 
now prevented Amstrad selling its cassette-recorders in precisely 
the form it had always intended, as Alan Sugar stressed after the 
result: 'We are not going to be bullied by powerful record and 
music industry pressure groups into withholding from consumers 
the advantages of developing technology.' 

For their part, the music companies had won a clear statement 
from Britain's highest court that the existing law on home taping 
was practically worthless. This statement could hardly have been 
better timed from their point of view, because it coincided with 
the passage through Parliament of the Government's attempt at a 
root-and-branch reform of copyright law in the Copyright Bill. 

At that stage the Government was committed to introducing a 



106 



ALAN sugar: the amstrad story 

levy on blank tapes - in effect, a tax on home taping. The music 
industry used the Law Lords' comments in the Amstrad case to 
lobby for similar action against manufacturers who launched 
recording machines which encouraged copyright infringement. In 
the end, however, the Government quietly laid to rest the notion 
of a blank tape levy - and, in effect, any parallel action against 
recorder manufacturers. Too many Conservative MPs rated the 
interests of music listeners well above those of the music com- 
panies. After almost four years of pursuing Amstrad, the music 
industry came away empty-handed. 



Alan Sugar's success with the Tower System, coupled with his 
swift moves in and out of markets, shone through Amstrad's 
financial results in the early 1980s. Profits increased by 75 per cent 
in 1980-81 (Amstrad's financial year ends in June); by 101 per 
cent in 1981-82; by 69 per cent in 1982-83; and by a relatively 
modest 13 per cent in 1983-84, when profit margins were squeezed 
by the venture into large-screen televisions and video recorders, as 
well as by the fall in the value of the pound which boosted the cost 
in sterling of Amstrad's imports of equipment and parts. Neverthe- 
less, figures for that year - profits of £9.1 million on turnover of 
£84.95 million - registered just how quickly the company had 
grown in the short time since its flotation. In 1984, Amstrad's 
progress received public recognition when the 37-year-old Sugar 
was named Guardian Young Businessman of the Year. 

The stock market was beginning to notice this odd little audio 
company. Amstrad's shares traded for 178p in October 1981, twice 
the price at which they had been floated. In February 1982, they 
touched 205p. By October 1982, when the doubling in profits for 
1981-82 was declared, the shares reached 395p, thanks to a 1 1 5p 
surge in one week. People who had held Amstrad stock since its 
flotation in April 1980 had seen the value of their investment 
increase four-and-a-half times in thirty months. 

So strongly were the shares performing that Amstrad carried 
out a one-for-one scrip issue in late 1982. This is an accounting 
device which halves the face value of a share and doubles the 



107 



DAVID THOMAS 

number of shares in existence. A scrip issue makes no difference to 
the value of individual shareholders' stakes: instead of holding 100 
shares worth 400p each, for example, they find themselves holding 
200 shares worth 200p each. But conventional wisdom in the City 
is that once the face value of a share becomes too high, some 
psychological resistance emerges to trading in that share. Amstrad 
shares resumed their onward march, prompting the company to 
decide on another sub-division of its shares in late 1983 - on this 
occasion, by five times. 

The upward journey of the Amstrad share price had been 
closely watched by Alan Sugar - naturally enough, since on 
flotation he had been left with 75 per cent of the shares. In March 
1982, he realized £2 million by selling almost 900,000 shares at 
229. 5p each, cutting his stake in Amstrad to 65.4 per cent in the 
process. Almost a year later, in February 1983, he raised another 
£4.6 million by selling 1.2 million shares at 382p each, leaving him 
with 58.9 per cent of Amstrad. 

Towards the end of 1982 the press began to quote the value of 
Sugar's remaining stake in Amstrad which, thanks to the share 
price rise, was pushing him into the league of Britain's super-rich. 
The Evening Standard calculated on 6 October 1982 that Sugar's 
stake was worth £23 million, including a £3 million increase on 
the day following the release of Amstrad's 1981-82 results. By the 
end of January 1983, the value of his stake in Amstrad had more 
than doubled to £48 million. 

Of course, these sums represented nothing more substantial 
than a paper fortune. Sugar would never be in a position to realize 
the entire face value of his holding in Amstrad by selling all his 
shares on the stock market: the very news that he intended to do 
so would result in a collapse of Amstrad's share price because 
investors would conclude that he no longer had confidence in the 
company. Yet by February 1983 he had raised about £8.6 million 
in total from three sales of tranches of his Amstrad shares. This 
money - and larger sums which were to follow later - formed the 
basis of the personal fortune which he held independently of the 
company. So financially secure was Sugar that he felt able to 
waive about £1 million of dividend payments due him from his 



108 



ALAN sugar: the amstrad story 

Amstrad stake between 1981 and 1983: given the then high tax 
rates on unearned income, it made more sense for him to keep this 
money within Amstrad to develop the company. 

The Sugars moved house again, completing the journey out of 
the East End which Alan had begun less than twenty years 
previously. Their new home was in a lane just around the corner 
from their old house in Chigwell Rise, but in a totally different 
league from even the detached homes of the small businessmen 
and professionals who inhabit the Rise. Now they were able to 
move into a millionaires' row, where half a dozen newly-built 
mansions cater for some of Britain's wealthiest people. Extensive 
grounds glimpsed behind the driveways, tennis courts, private 
gyms and swimming pools - these were the norm among the 
Sugars' new neighbours. The Sugars bought a handsome, L- 
shaped house in red brick, nestled round a central courtyard, 
complete with roof terrace and living rooms opening out on to a 
large rock garden in the middle of extensive lawns. 

The Sugars' house is simpler and more classical than some of 
its neighbours, which include a sprinkling of mock Tudor man- 
sions, but it was more than enough to stun Alan Sugar's family. 
Ann tried to prepare Daphne, Alan's sister, by telling her that 
their new home was big; but when Daphne and husband Harold 
went out to see it, they had to drive past it a couple of times before 
they dared to enter: 'We couldn't believe the size of the property. 
We'd never seen anything like that,' Daphne says. 

Fay and Nathan, Sugar's parents, shared this reaction. 'When 
my mother and father first went to see it, they were totally 
confused. They couldn't take it in. They just hadn't realized how 
big Alan had got,' recalls Daphne. 



City professionals were also slow to wake up to Amstrad, despite 
the advance in its share price. W. Greenwell, Amstrad's brokers, 
struggled to convince the many large insurance companies and 
pension funds which had ignored the flotation to start buying the 
stock. It was psychologically difficult for institutions which had 



109 



DAVID THOMAS 



squandered the opportunity of buying Amstrad shares at 85p to 
acquire them after they had passed the 150p or 200p mark. 

Paul Williams of Clerical Medical and General Life Assurance, 
the institution which became the biggest shareholder after Sugar 
at the time of the company's flotation, was keen to encourage a 
wider interest in the stock in the City: this would tend to accelerate 
the upward movement of Amstrad's share price, provided the 
company continued to perform well. But Williams found that 
institutional doubters cited Amstrad's very success as a reason for 
staying clear: 'Other institutions were always frightened they 
might be getting in at the peak. I was often told by brokers that it 
would be easier to sell Amstrad to clients when it slowed down to 
20 per cent growth, because no one could believe that 50 per cent 
or 100 per cent growth was sustainable.' 

Clerical Medical remained impressed by Amstrad's record and 
gradually bought more shares to the point where by May 1984 it 
held 12.8 per cent. But it was bothered by the fact that Greenwell 
was the only stockbrokers whose analysts regularly followed 
Amstrad immediately after its flotation. This meant that the one 
source of informed data and analysis on Amstrad was Greenwell, 
which as Amstrad's brokers was naturally partisan. 

In 1980 Greg Morgan had started work as an analyst for Gilbert 
Eliott, a small firm of stockbrokers later taken over by the Austrian 
bank, Girozentrale. Morgan phoned Clerical Medical (one of his 
clients in his previous firm) to ask whether the insurance company 
would do business with him at Gilbert Eliott. Clerical Medical 
pondered the matter and returned with a proposition: it would 
start channelling business Morgan's way provided Morgan began 
to follow Amstrad. 

Morgan knew nothing about Amstrad but, an assiduous man, 
he put some effort into researching the business: it would do 
nothing for his reputation if he began to recommend purchase of 
shares in a company which flopped eighteen months later. He 
travelled to Hull and lunched in Comet's board-room with 
Michael Hollingbery, who told him that out of all Comet's 
suppliers Sugar had about the best feel for the market. Morgan 
was impressed by Amstrad's line-up of City advisers: 'You would 



110 



ALAN sugar: the amstrad story 

not have associated a Hackney lad with Kleinwort Benson,' he 
comments. 

But the key factor in swinging Morgan around to Amstrad was 
a visit he paid to Sugar one afternoon in Tottenham. The meeting 
was scheduled to last ten minutes, but it stretched to 45. Sugar 
talked and talked - not to Morgan, but at him - about audio, 
television, car radios and all the other markets Amstrad was then 
in. He spun the Amstrad philosophy of flexibility to Morgan and 
the City analyst came away with an image of a thrusting young 
businessman who could cut through to the heart of his markets 
with blinding simplicity. He briefed his salesmen about the stock, 
and they began to recommend Amstrad to clients; but, like 
Greenwell, they found it hard work to interest the institutions. 

Part of the problem lay with Amstrad. From day one as 
chairman of a public company, Sugar made it plain that selling 
Amstrad's products was a better use of his time than selling the 
company to the City. Amstrad even began to hold its annual 
general meetings in Tottenham, first in its Garman Road head- 
quarters and then in a room at the Spurs football ground premises, 
rather than in a more conventional venue near the City. Greg 
Morgan used to trail up to these A.G.M.s where analysts and 
shareholders were regularly outnumbered by the company's board 
and advisers. It was all part of the Amstrad ethic of satisfying City 
requirements in the most economical and time-efficient manner 
possible. 

David Thomson, Greenwell's head of corporate finance, tried to 
persuade Sugar to do the normal round of institutional briefings. 
'But he wouldn't come out of his hole in Tottenham,' Thomson 
recalls. Not only did visiting the City run against Sugar's natural 
temperament, but it was also in his opinion a waste of time: 
Amstrad's share price would depend on his making a profit, which 
meant selling goods, not on being nice to the City. 

In Sugar's view, the height of wasted time is the business lunch, 
including lunches in the City where the square mile's best and 
brightest can quiz senior figures from industry about their plans 
and progress: 'They get on my bloody nerves actually. I get very 
fidgety sitting around at lunch for 2V2 hours. There are people 



111 



DAVID THOMAS 

who get to the office at 9.30, read the FT until 10.30, then they 
have tea and biscuits, and at 11.00 they do their bit of brilliance 
for the day. At 1 1.55 they down their pen and go out to lunch, roll 
back at 3.30 and have a final bout of brilliance and that's it,' he 
told a journalist in 1980. 

Despairing of ever tempting Sugar into the City, Greenwell 
began to arrange broker visits to Amstrad's headquarters in 
Tottenham, where he was visibly more relaxed among the card- 
board boxes talking about company products. There appeared to 
be a pay-off from these briefings for Sugar: Greg Morgan began to 
detect a pattern in which the visits - which heightened City 
interest in Amstrad - would be followed a week or so later by 
Sugar selling a parcel of his stock. But even these share-selling 
exercises threatened to cause a rift with the City. On one occasion, 
Amstrad's share price fell sharply between the time when Sugar 
briefed Greenwell about his plans to sell and the execution of the 
sale, suggesting that his intentions had been leaked - a classic 
example of insider trading. Sugar reacted angrily, convinced that 
the leak came from within Greenwell though the brokers vehe- 
mently denied this. 

Some years later Sugar was to say on Terry Wogan's television 
chat show that he may have been naive in his dealings with the 
City. As chairman of a new public company, he was warned in the 
clearest terms by the Stock Exchange not to divulge price-sensitive 
information selectively. Wanting to play scrupulously by the rules, 
he made a point of talking to as few people from the City as 
possible. 

Perhaps: but there is no doubt that this line of argument neatly 
coincided with his natural inclination to regard many City people 
as overpaid time-wasters. The upshot was that the signals emanat- 
ing from Amstrad were always sparse and sometimes misleading. 
Greg Morgan remembers being leaned on by the firm's City 
advisers not to be too enthusiastic in his profit forecasts: when 
Amstrad duly reported another set of stunning results, the market 
reacted to the news that they had greatly exceeded the City 
forecasts by marking up the share price sharply. After an unusually 
disappointing Christmas for Amstrad in 1984, by contrast, 



112- 



ALAN sugar: the amstrad story 

Morgan did not receive a hint of this news from the company until 
the early spring. 

Yet Sugar was developing his own form of communication with 
the outside world, including the City: the chairman's statement he 
released at the time of Amstrad's annual results. It is a near 
universal practice for chairmen of public companies to issue 
statements when releasing their results, but the accompanying 
figures usually contain the only serious points of interest. The 
typical chairman's statement is drafted by some bright young 
executive at corporate headquarters, then put through the sieve of 
the company's public relations and financial advisers until it 
emerges with all the excitement of pureed baby-food. 

By contrast, Sugar has always put his personal stamp on his 
chairman's statements. While obviously scrutinized by his advis- 
ers, they are written in a language and with a directness that 
could have emanated only from his own hand. As a result, they 
provide useful pointers for Amstrad watchers in the City and 
elsewhere. 

As early as October 1981, when Sugar was reporting on 
Amstrad's results for 1980-81, he penned an account of his 
strategy which amounts to the first sustained statement of the 
Amstrad philosophy. Unveiling a 75 per cent increase in profits, 
'during a period which has been described by some as a recession', 
he attributed these results to 'our expertise in marketing and our 
efficiency in production and management'. 

Sugar had seen no need to expand management or clerical staff 
despite a 61 per cent increase in sales: 'Our administration staff 
levels have remained constant and have coped with the company's 
expansion efficiently, and we remain confident that further expan- 
sion can be handled equally well.' 

Manufacturing flexibility was at the heart of Amstrad's strategy: 

The company has always used a flexible approach to production 
and we still wish to maintain the options of 100 per cent 
manufacturing or part assembly or indeed, importation and 
distribution of finished products. It is important to note that the 
recent dramatic devaluation of the pound would have left the 



113 



DAVID THOMAS 

company in a difficult position had we been committed to one 
way of producing. With our approach, we were able to turn off 
quickly certain items imported from the Far East and produce 
them in the U.K. In the past, when the reverse has been 
applicable, it was just as easy to hive off the making of some 
items to other countries. 

Revealing that Amstrad was working on a stream of new products, 
Sugar argued that, 'we have become trend-setters in design 
concepts, similar in a way to the major Japanese companies, but 
with the important difference of not being committed to any one 
method of production or area in which to produce.' 

The new products to be unveiled in the coming year would, 
Sugar assured Amstrad's shareholders, be 'in true Amstrad fash- 
ion, one step ahead of the market and most definitely the 
competition'. 



• 114 



7 



The People's Computer 



It was a baking hot day in early August 1983 and Roland Perry 
was becoming bored with the annual stock-take. He worked for a 
small company which traded as Ambit International and made its 
money by running a mail order catalogue for electronic compo- 
nents, a business that financed a range of other activities more to 
the taste of the electronics buffs who made up the company. They 
published an electronics magazine and designed electronic gadgets 
such as calculators. They had even helped Amstrad in the past by 
putting some finishing touches to its CB radio. 

Counting the number of transistors in stock did not rank as one 
of Perry's most exciting responsibilities. So it was with more than 
usual interest that he saw Bob Watkins walk into Ambit's high 
street premises in Brentwood - a small town in the middle of the 
Essex commuter belt, where the suburban sprawl to the north-east 
of London begins to give way to countryside. Amstrad's technical 
director knew the Ambit people from their past work for the 
company - work which had come Ambit's way because its 
managing director, William Poel, was a friend of Stan Randall, 
who by then was running Amstrad Hong Kong. 

Watkins was carrying a cardboard box which he placed on a 
desk in front of Perry. He produced from the box, first a wodge of 
papers and then a brand-new computer keyboard. Amstrad's 
name was unmistakable in a corner of the keyboard. 

'You know about computers,' Watkins said to Perry. 'Well, 



115 



DAVID THOMAS 



we've got this computer project we're working on and the designers 
have done a bunk. Do you think you could finish it off for us?' 

This opening gambit raised more questions than it answered, so 
Watkins elaborated on the reason for his journey to Brentwood. 
Throughout 1982, Alan Sugar had become increasingly impressed 
by the home computer boom spearheaded in Britain by companies 
such as Sinclair, Acorn, Commodore and Oric. They were selling 
tens of thousands of machines, many for £200 or £300 each - 
dwarfing the money to be made in Amstrad's audio business. 
Eventually, Sugar decided that home computers offered a market 
ripe for the Amstrad treatment. 

Sugar knew nothing about computer technology so he asked 
around, quickly learning how different the business was from hi-fi. 
If Amstrad intended to make a lot of money in computers, then it 
could not buy off-the-shelf technology as it had for its audio 
products; it would have to design its own computers from scratch. 

Amstrad could handle the exterior of its computer, the way it 
looked; Sugar already had some innovative ideas on that score. 
But it would have to rely on others to design the machine's interior 
— its electronic circuitry and operating software. For this, Sugar 
turned to a couple of engineers who had worked for Amstrad in 
the past and who claimed to understand computers. 

In the first half of 1983 Sugar and Watkins busied themselves 
finalizing the look of their computer. Ignorant of just how much 
effort goes into designing a computer, they naively assumed that 
their two engineers were making similar progress, but by the 
summer of 1983 they had begun to suspect that the circuitry and 
software design were seriously awry. These suspicions were con- 
firmed when the software designer cracked under the pressure, 
broke off contact with Amstrad and disappeared. 

Watkins concluded his story: C I managed to track down where 
he lived and get access to his place. I found bits of computer 
strewn all around the house, as well as all the designs for the 
Amstrad computer,' he said, nodding at the pile of papers in front 
of Perry. 'I also noticed a letter from Barclays Bank, so I contacted 
the bank which said that a cheque had been cashed in the Lake 



• 116 



ALAN sugar: the amstrad story 

District. I reckon he's disappeared to the Lake District, so at least 
we know he hasn't killed himself.' 

Asking Perry whether Ambit would finish the project for 
Amstrad, Watkins made only two stipulations. First, Amstrad's 
plans to enter the computer market had to remain secret: Ambit 
must tape over the Amstrad name whenever outsiders were shown 
the keyboard. Second, the project had to be completed fast - 
incredibly fast, since Amstrad wanted the finished design before 
the end of January, some five months away. 

Roland Perry talked over Amstrad's request with his managing 
director, William Poel, and they both agreed to work on the 
project. Their role would be as project managers: they would find 
the best designers and oversee the operation. Initially, the two 
Ambit people envisaged having to devote about half their time to 
Amstrad - a time allocation that in the event proved wildly 
optimistic. 

Short and wiry, Perry is the type of engineer who elevates 
method into the guiding principle of his work. His first move was 
to contact the man responsible for the hardware design thus far - 
the member of the original duo who had not disappeared to the 
Lake District. The meeting, held in a pub, was not promising, as 
Perry recalls: 'He was an oldish guy, bearded, with a moustache 
and smoking a pipe. He arrived in the pub with his young son, 
who was about 13 or 14. As the conversation progressed, it became 
clear that the son had been doing most of the work on the Amstrad 
computer.' 

Perry was unimpressed with the work of this man-and-boy 
partnership. Although their computer was able to put a few letters 
on the screen, the whole design was untidy and inefficient; the 
colour graphics in particular were a disaster. On leaving the pub, 
he decided to assemble a new team. 

The newly appointed project manager set about tapping a 
network of contacts dating from his days as an engineering student 
at Cambridge University. Cambridge is the nearest Britain has to 
Silicon Valley, that legendary strip of California which has 
spawned many breakthroughs in computing and electronics; the 
university's catalytic role in fostering a local high-tech economy 



117- 



DAVID THOMAS 

was similar to that played by Stanford University in Silicon Valley. 
From the 1960s onwards, an easy interchange developed between 
the academics in Cambridge's world-class science and engineering 
departments and the electronics companies that were starting up 
all over town. Many key figures who shaped the Cambridge 
phenomenon divided their time between the University and their 
entrepreneurial activities. By 1983 computing, biotechnology, 
lasers, advanced medical equipment and other high-tech industries 
accounted for almost one in six of the workers in Cambridge and 
the surrounding region - an area dubbed 'Silicon Fen' by the 
media. Sinclair Research and Acorn were the acknowledged stars 
among a host of other companies which shone in the Cambridge 
firmament. 

Perry was at Cambridge in the early 1970s with a group of 
friends who saw the beginnings of this intellectual and entrepre- 
neurial excitement. Like him, many of his friends had been 
inspired by the maths and physics teaching at Brentwood School 
to study engineering or maths at Cambridge. The Brentwood gang 
merged naturally into a wider circle of like-minded techies at the 
university. Many went on to work in computing. It was to this 
circle that Perry now turned. 

Perry moved quickly, travelling to Cambridge within days of 
Watkins's visit to talk to two sets of computer designers. The most 
pressing problem was to find software designers for the Amstrad 
computer. It not only needed an operating system - the master 
program which controls the main activities of a computer - but 
also a Basic, the most widely used programming language in 
personal computers. Only when the operating system and Basic 
were fixed would software houses be able to write games and other 
programs which could be used on the Amstrad machine. 

The most popular version of Basic is that written by Microsoft, 
the American company which is one of the world's largest software 
concerns. But Amstrad would have had to pay a large sum for the 
right to use Microsoft's Basic, so Sugar had decided that his 
computer would have its own specially written Basic. 'Save money 
by designing from scratch' was to become a maxim for Amstrad's 
early approach to computers. The trouble was that the company's 



118 



ALAN sugar: the amstrad story 

original software designer had made virtually no progress before 
disappearing to the Lake District. 

One design firm in Cambridge gave Perry a dusty reception, 
irritated by his refusal to say anything about his clients other 
than that they were 'a big high street company'. But he struck 
lucky with another contact, Howard Fisher, also ex-Brentwood 
School and Cambridge. Fisher had recently managed a project 
for Acorn, the Cambridge-based computer company, which 
involved developing a Basic for a new Acorn small business 
package. The work had been done by a two-man design team, 
Chris Hall and Richard Clayton, who had recently formed a 
partnership named Locomotive Software. Fisher suggested to 
Perry that he should get in touch with Clayton, who was part of 
the old Cambridge gang. 

Perry rang Clayton the next day, fixed a meeting and drove 
down to Dorking, a town in stockbroker-belt Surrey, parking 
outside Clayton's house. Locomotive Software - then barely six 
months old - operated out of Clayton's back room. Richard 
Clayton is the nearest British equivalent to the laid-back software 
wizards of California. Still sporting the long hair of a 1960s hippie, 
he talks with infectious enthusiasm whenever the conversation 
turns to software. 

Roland Perry sat on a sofa in Clayton's house in August 1983 
and showed the Locomotive duo the keyboard with Amstrad's 
name taped over. The first thing Clayton did was to fetch a 
screwdriver and open up the keyboard, spotting in a corner of the 
circuitry an uncovered Amstrad label. But the discovery added 
little to his knowledge: 'I didn't actually know who Amstrad were, 
so it didn't mean anything to me,' he recalls. 

The fact that Perry's mystery client had already completed the 
outside of the machine was a plus point with Locomotive: 'The 
thing that really impressed me was they'd done the plastics. One 
of the things I'd learnt was it often took 18 months to get the 
plastics right. These people had already done the box. It was 
incredible. They were clearly serious,' Clayton says. 

However, it was obvious from Perry's account that the software 
side of the project had stalled. The electronic circuitry also looked 



119 



DAVID THOMAS 

ropey, although neither Hall nor Clayton regarded themselves as 
hardware experts. Clayton told Perry that Mark-Eric Jones, a 
hardware expert and another member of the Cambridge gang, 
lived only ten miles away. 'Why don't I give him a call and see if 
he can come over?' suggested Clayton. 

An engineer who chooses his words carefully, Mark-Eric Jones 
is universally known by his initials, Mej (rhyming with wedge). 
Mej's father founded and ran a computer company called Data 
Recall, which employed many of the ex-Cambridge circle in the 
1970s and early 1980s, including Richard Clayton. Data Recall's 
star product was a word processor, the Diamond, which was well 
regarded in the days when word processors cost about £10,000 
each. It was a background which was later to serve Amstrad well. 

Mej had set himself up with a friend as an independent 
electronics designer, going under the name of MEJ Electronics. 
When the two electronic engineers came over to Clayton's house 
in Dorking, they needed only one look at the interior of Perry's 
machine to start poking fun at it: 'Oh, that's interesting, this 
power supply isn't linked up to half the chips. That's why it 
doesn't work,' said Mej. 

Perry wanted to press the designers into a decision. Top of the 
agenda was the issue which had been hanging over the talks all 
afternoon - time. 'How long would it take you to do an operating 
system and a Basic for a 6502 processor?' he asked the Locomotive 
pair. 

'About eight months,' Clayton replied. T reckon we could 
deliver it by April.' 

The answer alarmed Perry, since Amstrad needed prototypes of 
the machine by December. The essential problem had become 
clear during the course of the afternoon. Neither Locomotive nor 
Mej had had much experience with the particular microprocessor 
which Amstrad envisaged for its machine. Microprocessors are the 
brains of a personal computer, the chips which control its core 
operations. Each one - usually known by a number given to it 
by its manufacturer - has its own particular architecture for 
executing instructions. Amstrad had chosen the Motorola 6502 
because it was a standard processor in many of the computers 



• 120 



ALAN sugar: the amstrad story 

with which it was likely to compete. Many games and other 
programs designed for 6502-based machines were already on the 
market: a computer company needs to be able to point to such a 
library of end user programs if it is to sell its machine in any 
numbers. 

Clayton added almost as an afterthought: 'If you did it with the 
Z80, you could have the software much earlier.' 

This was what Perry had been waiting to hear, so Clayton 
developed the point. Locomotive could complete one half of the 
software job, the Basic, for a Z80 microprocessor more quickly 
because it had already done one. The Acorn machine whose Basic 
Clayton and Hall had just written was based on the Z80, so they 
would be able to adapt their Acorn work. Moreover, Mej also 
knew the chip back to front, because Data Recall's word processors 
had been based on Z80 chips. 

As afternoon turned into evening, everyone moved from Clay- 
ton's house to a local pub for a drink and a meal. It was in the 
King's Arms, Dorking, that the final design for Amstrad's first 
computer was hammered out. They would throw away the existing 
software and hardware design, transform the computer into a Z80- 
based machine and knock out an operating system and a Basic by 
the end of January. At the end of the evening, the Dorking 
engineers remained in the dark about the client's identity. Perry 
had talked all day about 'a big high street name'. Clayton went 
home thinking: 'Dixons? What is Dixons doing building a personal 
computer?' 

Shortly afterwards, Perry solved the problem of how to keep his 
client's name secret. It was pointless to continue taping over the 
Amstrad name, because someone was bound to remove the tape. 
Since he had been telling everyone that his client was 'a very well 
known British electronics company', he called the machine by the 
codename Arnold, replacing 'Amstrad' on the prototype keyboards 
with 'Arnold'. Most people assumed that his client was the 
General Electric Company, Britain's biggest manufacturer which 
is headed by Lord (Arnold) Weinstock. Perry did nothing to 
discourage this speculation; only later did he notice that Arnold is 
also an anagram of his own first name, Roland. The computer was 



121 



DAVID THOMAS 



later to acquire the formal name of CPC464 - C(olour) P(ersonal) 
C(omputer), coupled with 64 as an indication of its memory size. 

Locomotive and Mej had waved goodbye to Perry, more than 
half expecting never to hear of the project again: 'We didn't really 
give much chance of him coming back to us, because our proposal 
was so far away from what Roland had turned up wanting/ 
Clayton says. 

But they had not reckoned on Amstrad's wish for speed. 
Amstrad was impressed with how quickly the Dorking designers 
promised to deliver the machine if the microprocessor was 
changed. A few days later, Locomotive and MEJ Electronics 
received a call giving them directions to go to the mystery client's 
headquarters in North London. 

Chris Hall represented Locomotive, because he alone of the 
Locomotive pair possessed a suit. Mej was there too, along with 
William Poel from Ambit and Bob Watkins from Amstrad. Alan 
Sugar came in late, having just stepped off a plane at Heathrow 
airport. No one introduced Sugar to Mej, who spent part of the 
meeting wondering about the identity of the latecomer. Initially 
he thought that Bob Watkins must be Amstrad's boss, until he 
noticed that every time the latest arrival spoke everyone else 
stopped talking. 

Sugar explained his vision of Amstrad's entry into the com- 
puter market. He was going to produce a computer, he told the 
assembled electronics buffs, that had 'perceived value for money'. 
Although he was proposing to launch what was essentially a 
games machine, it had to look like the computers people saw 7 in 
offices and airports. That was one reason why Amstrad was 
planning a machine with a large keyboard and screen. It also 
partly explained why Amstrad was proposing to do something 
almost unknown in low end home computers: to produce a 
machine with a built-in monitor and a built-in cassette recorder 
for storing programs. 

Sugar had devised this plan not long after he took an interest in 
the computer market. Having bought some of the machines then 
on sale to teach himself about the technology, he was amazed how 
difficult it was to use most home computers. Since they did not 



122- 



ALAN sugar: the amstrad story 

come complete with their own monitors or cassette recorders, users 
had to link up the computer with a television set and tape recorder. 
This messy process caused no end of aggravation as technically 
illiterate people struggled to connect the different machines. The 
sound and picture quality which resulted were often atrocious. 

In short, the home computer market was crying out for the 
clarity which Sugar had brought to the audio industry with his 
Tower System. As he explains: 

Our first computer was a very typical Amstrad concept. We sat 
down and observed all this computer stuff. And we saw what 
people were actually buying with computers. They needed lots 
of cables and cassette decks, and then they had to plug it into a 
television. There were lots of rejects coming back because of 
mismatches on products. I decided that the Amstrad philosophy 
is an all-in-one piece, so we would present our product as 
complete with a keyboard, cassette mechanism and monitor. 

The one-piece solution would end family feuds over access to the 
television set caused by the home computer craze. Since most 
parents were reluctant to surrender the family TV for their 
children's new toy, many home computers were discarded as 
useless junk once the novelty had worn off. Sugar saw the need for 
a home computer that did not rely on the family TV: 'It was 
obvious that you had to keep the computer far away from the 
main television in the home. And because of all the wires and 
other gubbins you had to struggle with, many Sinclair computers 
were ending up thrown under the bed, never to be used,' Sugar 
recalls, adding, 'Our computer was too big to be thrown under the 
bed.' 

The size of the computer was also central to Sugar's marketing 
plans: 

It looked like a mug's eyeful for the old man when he walked 
into Dixons. He looks at this thing, with its whacking great big 
keyboard and a monitor, and he has visions of a girl at Gatwick 
airport where he checks himself in for his holidays. And he 



123- 



DAVID THOMAS 

thinks, 'That's a real computer, not this pregnant calculator 
thing over there called a Sinclair.' So that was my marketing 
concept: the old man, who has got to fork out a couple of 
hundred quid on kit because the kid is driving him mad for a 
computer, sees this thing that actually looks like a real 
computer. 

The one-plug, all-in-one computer allowed Amstrad to repeat the 
component savings achieved with its Tower System. Duplicated 
power supplies, transformers and so on were cut out. This was 
fundamental to Sugar's ability to keep the price of his first 
computer low. Further cost saving was achieved by the canny 
choice of components. He used television tubes rather than the 
more specialist computer visual display unit tubes in the CPC464 
because they were cheaper. He also came across a supply of 3" 
disks and disk drive mechanisms in Japan, which were being sold 
cheaply because the 3" format was in the closing stages of an 
ultimately losing battle against ZVi' disks. Amstrad's computers 
were to be the last strong-selling line to use 3" disks. 

At the initial meeting with his new designers, Sugar refused to 
tell them the price he had in mind for the machine. Indeed, he 
became quite angry when they asked. The most he would disclose 
was a general indication of the computers he intended to compete 
against. He was determined to keep his pricing plans secret, but 
his reticence also reflected his belief that giving the designers a 
figure would simply encourage them to make the computer that 
expensive. 'Do it as cheaply as is humanly possible' - this was the 
message. 

Otherwise, the specification given by Amstrad to the designers 
was about as minimal as can be imagined. The machine had to 
have colour, sound and a memory capacity of 64k (64,000 pieces 
of information) for users, as well as 32k of ROM (read only 
memory, on which the computer's main operating programs were 
stored). Sugar had identified 64k memory as essential to his 
machine's success, since this was the maximum amount of memory 
available in most other home computers. Matching that was as 
important to him as making sure that the watts potential of his 



124 



ALAN sugar: the amstrad story 

audio equipment was up with the leaders. He thought that 
computer-illiterate people would seize on size of memory as the 
one feature to check when they walked into a shop. 

Chris Hall and Mej left this initial meeting with mixed feelings. 
Unlike his friends, Hall had heard of Amstrad. His brother had 
once bought an Amstrad cassette recorder and they had decided 
between them never to buy anything with the Amstrad name on it 
again. But their unease about working for a company serving the 
low end of the audio market was outweighed by the excitement of 
the project. After all, what had they to lose? Neither set of 
designers was so overwhelmed with work that they could pass over 
the chance to design a personal computer from scratch, an 
opportunity given to very few of the world's electronic designers. 
Mej also came away impressed by the sharpness of Sugar's vision: 
'He was the exact opposite of some people who just muddle 
around. He had this immense clarity in his ideas and in what he 
was trying to achieve.' 

Moreover, the designers were attracted by the sheer absurdity 
of what was being asked of them. When I.B.M. designed its first 
personal computer, admittedly a more complex machine, it gave 
26 in-house engineers, supported by 100 manufacturing tech- 
nicians and an array of sub-contractors, a year to complete the 
project - and this went down as a miracle of speed in the annals of 
the world's biggest computer company. Amstrad was giving half a 
dozen engineers barely five months to design its first computer. 'It 
was sufficiently close to impossible to do in that time scale that it 
was a very exciting challenge,' Mej remembers. 

Both Locomotive Software and MEJ Electronics expected a 
temporary involvement with Amstrad, because neither was opti- 
mistic about Amstrad's computer. 'Acorn's B.B.C. machine was 
going great guns at the time and it seemed to us that we were 
unlikely in the time scale available to be able to compete on a 
techie level either with that machine or with the latest Commo- 
dore. And Commodore was known to be working on some further 
machines, so it was difficult for us to see quite how this was going 
to sell,' Hall explains. 



125 



DAVID THOMAS 

The designers' pessimism coloured their approach to a follow- 
up meeting with Amstrad to negotiate their fees. Mej agreed a 
fixed price for his hardware work, preferring that to a royalty 
payment because of his scepticism about the number of machines 
that would sell. Locomotive quickly settled on a similar deal for 
the computer's operating system, but since it wanted to keep the 
intellectual property rights to its Basic, Chris Hall initially quoted 
£30,000 a year for the Basic licence to Bob Watkins. Eager for a 
substantial immediate payment, Locomotive actually offered 
Amstrad an even better deal - the use of the Basic for the first two 
years in return for £45,000 cash down. 'And what about the 
subsequent years?' Watkins asked. 

Hall did not think there would be any subsequent years for the 
machine, so he settled on a figure of £15,000 a year thereafter. Six 
years later, at the end of the 1980s, the CPC464 was still selling 
and Locomotive was still receiving a measly £15,000 each year 
from Amstrad for use of the Basic. 'How naive we were. Never 
mind. It seemed a lot of money at the time,' Hall says, laughing 
weakly. 

During these early stages, Mej pressed on Amstrad a change in 
design that was to have immense significance for its subsequent 
computer lines. He suggested that they should use gate arrays. 
Originally developed by Ferranti, gate arrays allow a large number 
of discrete components to be packed on to a silicon chip. They also 
allow a chip to be custom-designed so as to meet a manufacturer's 
precise specifications for a particular piece of equipment. Back in 
1983, gate arrays - also known as uncommitted logic arrays - were 
a relatively new technology for mass market consumer items; not 
even the mighty I.B.M. had used them in its first personal 
computer, although they did feature in Sinclair's home computers. 

Mej told Sugar that gate arrays had three advantages. First, by 
packing many components on to a chip, they would cut the 
component count and cost of his machine. Second, they would 
allow more features to be built into the computer for the same 
price. Third, they would reduce the chances of the Amstrad 
machine being copied, because once the customized gate array has 
been made its circuitry is effectively hidden. Sugar was initially 



126 



ALAN sugar: the amstrad story 

reluctant to use this novel technology, because he was worried 
that the designers would need more time to grapple with it. 
However, he rapidly grasped the potential of gate arrays. He was 
particularly impressed by their anti-copying properties, since he 
had seen enough LB.M. and Apple clones in the Far East to worry 
that his machine might be copied. Typically, it took Sugar just 
one meeting to make up his mind: go for gate arrays. After all, 
they replicated in miniature the Amstrad philosophy of paring 
cost by cutting out surplus components. 

Roland Perry telexed details of the final specification to Bob 
Watkins for his approval on 18 August 1983. The ROMs embody- 
ing the computer's operating system were due to be sent by air to 
Orion, Amstrad's Japanese sub-contractor, in the third week in 
January 1984. But the time scale was in reality even tighter than 
that, because of the importance of having games ready to use on 
the computer the very day it was launched. It would take some 
time for software houses to write games tailor-made for the 
Amstrad machine, so Sugar decided to ask 50 software companies 
to supply one game apiece. Each of these software houses would 
need a prototype machine, which meant despatching 50 prototypes 
by late November. 

Mej and his partner stocked up on frozen pizzas to sustain them 
late at night as they worked on the circuit designs in Mej's spare 
bedroom. They used pencil, paper and a drawing board: the 
computer-aided design techniques which were to become standard 
among electronic designers by the end of the 1980s were then way 
beyond their financial reach. Locomotive Software expanded to 
four people and moved into a tiny office in nearby Leatherhead. 
When the pace approached breaking point, the designers consoled 
themselves by joking that they could always disappear to the Lake 
District. 

Amstrad left Locomotive and Mej remarkably free from interfer- 
ence. No one at Amstrad could have contributed much to the 
technical design of a computer at that stage. Bob Watkins, the 
person most likely to have become involved, was fully stretched 
during the second half of 1983 with a production problem on the 
large-screen television line at Amstrad's Shoeburyness factory. 



127 



DAVID THOMAS 



Communications between Amstrad and its project managers in 
Ambit were not helped by the different cultures to which the two 
sets of people belonged - an issue which assumed greater import- 
ance after Amstrad's successful launch into computers. Amstrad 
worked rigid 9 am-5.15 pm hours, partly as a reflection of Sugar's 
abiding belief that efficient time-management allows any amount 
of work to be crammed into a standard working day, but also 
because of the constraints imposed by Amstrad's headquarters in 
Tottenham. Security was of paramount importance at Tottenham, 
since the offices fronted a large warehouse stacked high with 
Amstrad products. Every afternoon at 5.05 pm, the security guard 
would warn the Amstrad staff that he was locking up in five 
minutes. Everyone from the chairman downwards had to drop 
their work and clear out of the building, on pain of being locked in 
for the night. 

By contrast, Ambit worked the more relaxed hours of computer 
buffs: 10 am-1 pm and 3 pm-8 pm. As a result, Perry recalls: 'We 
found Amstrad incredibly difficult to do business with. In effect, 
we only had a short period in the afternoon when we could talk to 
Amstrad, and they tended to be very nervous in the last ten 
minutes about being locked in.' 

Yet Amstrad steered clear of detailed intervention in the design 
process, even during later computer projects when Sugar and 
Watkins had a better grasp of the technology and there were no 
communications constraints. The company believed that its 
designers had to be trusted to complete the job and that constant 
changes to the specification served only to slow down a project. 
MEJ Electronics had more contact with Amstrad than Locomotive 
Software during most of the projects, partly because MEJ's 
hardware designs had greater cost implications than Locomotive's 
software work. Also, with its background in the audio industry, 
Amstrad had a stronger instinctive feel for components than for 
software. 

In its first flush of enthusiasm for the computer market and in 
order to save costs, Amstrad had forward ordered components in 
huge numbers, typically 100,000 each. Mej thus came under 
pressure to design in these forward-ordered components, even 



128 



ALAN sugar: the amstrad story 

though they were not best suited for his new design. However, 
Amstrad minimized this problem by swapping components it had 
ordered for parts which Mej wanted, and by disposing of un- 
needed parts in the then buoyant component market. 

As project manager, Roland Perry was MEJ's and Locomotive's 
day-to-day contact with Amstrad as they worked on the CPC464. 
He was the main communications channel, translating Amstrad's 
concerns into language the designers could understand and vice 
versa. An electronics engineer himself, Perry appreciated the 
virtues of the free hand which Amstrad gave its designers: 'Unlike 
with other companies, we didn't have people ignorant of what they 
were talking about deciding it should have a big red knob on the 
side.' 

Perry kept Amstrad happy by organizing staged demonstrations 
during the design phase. After Amstrad's abortive experience with 
its previous designers, Bob Watkins felt that Sugar might also 
need some independent reassurance that the new team knew its 
job. Guy Kewney, a leading computer journalist, was brought to 
Amstrad's Tottenham headquarters to test one of the first proto- 
types, on condition that what he saw remained confidential. 
Everyone's nerves were soothed when Kewney emerged impressed 
by the machine. 

Locomotive met Sugar only once during the development of the 
CPC464, towards the end of November after they had produced a 
functioning prototype. Chris Hall put on his suit again and went 
to Tottenham one Friday to take Sugar through the prototype. 
Sugar immediately noticed that the on-screen cursor blob could 
not be moved when the computer was first switched on; the 
CPC464 had been programmed so that the cursor would move 
only once a game had been loaded or some text was in the machine 
for editing. 

'But everybody will think the computer's broken because when 
you press the cursor key nothing happens,' Sugar protested. 

'Moving the cursor key at this point has no meaning. There's 
nothing for it to do,' Hall countered. 

'Well, it makes sense to me,' Sugar replied. 

The exchange was typical of the different perspectives of 



VB 



DAVID THOMAS 

Amstrad and its designers. Sugar was imagining the reactions of a 
customer walking into a shop not knowing the first thing about 
computers. Locomotive was thinking of what was logical from a 
programming viewpoint. This was one of Sugar's very few inter- 
ventions in the software design for the CPC464; it made the people 
at Locomotive shake their heads at the irrationality of non- 
technical people, but it was also obviously justified. 

In early December Roland Perry organized the distribution of 
prototype machines to the software houses which Amstrad hoped 
would write games for its computer. A feat of small-scale produc- 
tion was needed to satisfy his demand for fifty prototypes. Drawing 
on Ambit's contacts in the component industry, Perry arranged 
the manufacture of fifty circuit boards. He oversaw the production 
of a board with dozens of chips on it to simulate the gate arrays, 
which were not ready at that point. Perry collected fifty colour 
televisions from Amstrad which were used as substitutes for the 
monitors, and arranged for a local company to make fifty power 
supplies for the machines. 

William Poel visited the software companies, persuading them 
to sign confidentiality agreements committing them not to disclose 
Amstrad's name and then trying to enthuse them to write games 
for this company unknown to the computer industry. Bob Watkins 
sorted out the commercial side, negotiating the rates Amstrad 
would pay for the games. He found it an uphill task; many of the 
bigger games houses, in particular, took the attitude: 'We've got 
plenty of work for Sinclair and Commodore. Who is Amstrad?' 

Roland Perry employed a couple of young women, whom he 
called his 'angels', to drive the prototypes at high speed to games 
houses round the country. Alan Sugar took to ringing up Ambit 
regularly to check the exact location of each machine. If nego- 
tiations with a particular software house were heading nowhere, 
then one of Perry's 'angels' would be despatched at a moment's 
notice to snatch back the prototype. 

Meanwhile, the gate arrays were turning into a nightmare. 
Amstrad had asked various semiconductor companies to quote for 
the twin task of helping Mej to design the gate arrays and then 
manufacturing them. The contract went to Ferranti, because its 



130- 



ALAN sugar: the amstrad story 

quote was much lower than that of its rivals, but relations between 
them rapidly deteriorated thanks to what Mej describes as 4 a saga 
of disasters' at the Ferranti end. Ferranti's design work was riddled 
with errors. Mej had to fly to the north-west of England so 
regularly to visit Ferranti's chip centre that the taxi-drivers at 
Manchester airport began to recognize him. In the end, Amstrad 
was forced to turn its secret weapon on Ferranti - Alan Sugar, 
who phoned to tell the electronics company in one-syllable Anglo- 
Saxon words precisely what Amstrad thought of its performance. 

The experience did little for Sugar's already low opinion of 
British manufacturing. Deciding it would need a second source for 
the gate arrays, in order to protect itself against Ferranti, Amstrad 
gave the job to SGS, the Italian chip manufacturer later to merge 
with Thomson of France. Mej added trips to Milan to his regular 
round of visits to Manchester. 

The prototypes were needed not just for the games houses but 
also to show Orion, Amstrad's Japanese suppliers for whom it was 
keen to lay on an extensive briefing. The Japanese company had 
manufactured many products for Amstrad in the past, but this 
time Orion was being asked to make something completely outside 
its - and Amstrad's - experience. Amstrad's Japanese partners 
had to be assured that the British company knew what it was 
doing. 

Mej, Roland Perry and Bob Watkins boarded a plane for Osaka 
in December 1983, clutching a prototype computer which they 
carried to Fukui, a town on Japan's northern coast. Like most of 
Amstrad's subsequent computers, the CPC464 was destined to be 
made in an Orion plant in South Korea, but Fukui was the base 
for Orion's manufacturing headquarters and home to its chief 
engineers. Watkins busied himself with grinding, day-long nego- 
tiations over the price Amstrad would pay Orion to make the 
machine. Mej, for his part, was engaged in the novel experience of 
explaining a piece of equipment to Japanese engineers. 

Mej had to become accustomed to never dealing with a single 
Japanese engineer at a time. The Japanese operated in teams 
because they had an engineer specializing in almost every part in 
the machine. Baffled by his attempt to discuss the whole machine, 



131 



DAVID THOMAS 

at first Mej's Japanese counterparts could not take such a gener- 
alist engineer seriously. They began to treat Mej and Roland Perry 
with more respect only after Mej had suggested an improvement 
to the monitor- which was very much in Orion's area of expertise. 
Orion's engineers then immersed themselves in the task of discuss- 
ing modifications to the computer. But on the occasions when the 
English pair proposed some addition, the Japanese would say in 
unison 'Cost-up'. It was their way of indicating that the change 
would add to the machine's price: 'cost-up' entered the language 
of the Dorking engineers, along with the ever-present possibility of 
'disappearing to the Lake District'. 

The Amstrad team returned from Japan with an Orion request 
to change the screen colouring. The Japanese wanted Amstrad to 
use yellow letters on a blue background, the strongest possible 
contrast, because they were worried that the previous colour 
arrangements were indistinct. Amstrad agreed and Locomotive set 
about adapting the colour codes. The finishing touches to the 
design went smoothly thereafter, although the Dorking engineers 
still had to work round the clock. Locomotive's Chris Hall allowed 
himself Christmas Day off, when he took home the now nearly 
completed machine for his mother to play with. 

Locomotive met its target of shipping the master codes for the 
computer's operating program on the third Saturday of January, 
1984. First thing on the following Monday morning, Richard 
Clayton discovered a bug. It was not a major bug - it affected a 
function that would be used only rarely and Locomotive knew how 
to fix it by the afternoon - but it was a bug all the same. So 
Clayton rang Amstrad, explained the problem and said: 'This is a 
trivial thing, we can change the codes because they can't have 
reached Orion's factory yet.' 

But Amstrad refused to ask Orion to change the master codes. 
The company was not prepared to alter something which had 
already been shipped in final form to Japan, and decided instead 
to remove the faulty function in question from the computer 
manual. This was just one of many examples down the years of 
Amstrad's determination never to lose face with its Japanese 
suppliers: to out-Japanese the Japanese at all times. 

• 132- 



ALAN sugar: the amstrad story 

Alan Sugar saw his first computer essentially as a games-playing 
machine for the home. But, unusually for that time, he also wanted 
it to have one foot in the business market. Part of the plan behind 
making the CPC464 big enough to look like an office computer 
was so that users would realize its potential for business functions. 
The CPC's screen was wide enough to take 80 characters, similar 
to most computers found in the office and twice that normally 
available for a home computer. 

Initially Amstrad had decided against producing a version of 
the CPC464 with disk drives which could support business pro- 
grams, but in March it reversed that decision and told its engineers 
to start work on a disk drive unit which could be added on to the 
CPC464. Amstrad then searched around for some business pro- 
grams to run on the machine. Digital Research, one of the best- 
known software companies in the United States, was a logical 
place for Amstrad to look: one of its flagship products, the CP/M 
operating system, had been working on Z80-based machines since 
the chip was first introduced. Providing a version of CP/M for the 
CPC464 would allow business applications like spreadsheets and 
word processing to be carried out on the new computer. Amstrad 
quickly came across Paul Bailey, then hard at work developing 
Digital Research's European sales. 

Bailey is a sharp, fast-talking salesman who has spent so long 
peddling the products of the American software industry through- 
out Europe that he has come to speak exactly like an American 
software engineer. After spending the early and mid 1980s building 
up Digital Research in Europe, in 1988 he moved to Lotus, one of 
Digital's arch rivals, and proceeded to do the same job for them. 

Bailey was more than happy to sell Amstrad the CP/M operat- 
ing system. Digital Research was then engaged in an ultimately 
losing battle against the rival MS-DOS operating system produced 
by Microsoft: MS-DOS had the overwhelming advantage that it 
had been adopted by I.B.M. for its first personal computer. Bailey 
was in the forefront of that battle, trying to sell CP/M to European 
computer manufacturers. Unlike Locomotive Software, Digital 
Research was as professional as Amstrad in commercial nego- 
tiations. Amstrad agreed to pay a set price in advance for a given 



133 



DAVID THOMAS 

volume of copies of the Digital software: once the deal was struck, 
Amstrad was committed to paying Digital Research the money no 
matter how many of its machines were sold. 

Bailey describes his CPC464 deal with Amstrad: 'Sugar basi- 
cally attracted me by agreeing to very large volumes in advance. 
The drawback was that he insisted on a very keen price. But in 
the event, we ended up selling over a million copies of CP/M to 
Amstrad for the CPC and their later word processors. So it 
certainly wasn't a low runner.' 

Amstrad's first computer, the CPC464, was unveiled in April 
1984 - a mere eight months after Bob Watkins had walked into 
Ambit's offices. By then Amstrad had hired public relations 
advisers, Michael Joyce Consultants, who were to mastermind its 
increasingly slick product launches. The great hall of Westminster 
School, a stone's throw from the Houses of Parliament, was hired 
for the CPC464's launch. Archimedes, Einstein, Monet and Wil- 
liam Shakespeare were just some of the celebrities on hand to 
usher Amstrad's first computer into the world: improbably, Nick 
Hewer of Michael Joyce had managed to find people living in 
Britain who were blessed with these names. Monet showed the 
fifty or so computer pundits the computer's colour capabilities, 
Shakespeare talked through its word processing facility and so on. 

The CPC464 came in three formats: a green on black screen for 
£229, colour capability version for £329 and a version with the 
add-on disk drive for £429. It emerged to a generally good 
reception in the press, with some journalists dubbing it 'the 
people's computer'. Jack Schofield, computer editor of the Guard- 
ian, wrote: Tt out-performs the Commodore 64 (the machine's 
main rival) in virtually every department', adding that it was 
extremely cheap, costing £80 less than the equivalent Commodore 
or Atari machine. 

Guy Kewney, the computer journalist who had had an advance 
sight of the CPC464, praised its all-in-one design as a break- 
through which could finally rid the industry of its 'enthusiast' 
image. His four-page bench-test of Amstrad's computer for Personal 
Computer World concluded: 'The Amstrad is a powerful, fast 
machine, with plenty of memory, easy to program, and packaged 



134 



ALAN sugar: the amstrad story 

in a way that means it will comfortably outsell the Acorn Electron, 
and give the Commodore 64 and Sinclair Spectrum a hard run for 
their money. I expect some 200,000 systems to be sold by the end 
of the year.' 

Not everyone was so optimistic. 



When Alan Sugar first became interested in the home computer 
market back in 1982, it seemed ripe for the Amstrad treatment. 
Home computers were the consumer electronics phenomenon of 
the early 1980s. Dazzling fortunes were made as a clutch of 
entrepreneurial companies like Sinclair Research and Acorn began 
to count their sales in hundreds of thousands of units. Amstrad 
appeared poised to carry through what it was best at doing: 
learning from the mistakes of the existing players to produce a 
value-for-money machine which left the competition standing. 

Britain's fascination for personal computers followed in the 
wake of the boom in the United States, where this precocious 
industry was born. Dating the birth of an industry is an inexact 
science, but many people would plump for the day in 1977 when 
Apple Computer launched the Apple II, the first personal com- 
puter which users could put to work without the need for assemb- 
ling or programming. Apple was the brainchild of two men — Steve 
Wozniak, a self-taught engineer, and Steve Jobs, hippie turned 
high tech entrepreneur. 

Jobs supplied the vision which launched the personal computer 
revolution; he wanted to dethrone the Big Brother image of 
computers that was familiar to anyone who watched a science 
fiction movie in the 1950s and 1960s. Computers were seen as 
huge cabinets with whirring reels of tape whose mysteries were 
guarded by a new priest class - data processing managers. Jobs 
wanted to empower the people by putting a computer on every 
desk. 'We build a device that gives people the same power over 
information that large corporations and the government have had 
over people,' an early Apple video proclaimed. 

This message chimed perfectly with the counter-culture ethos of 



135 



DAVID THOMAS 

many computer freaks in Apple's native California. More import- 
ant, it also coincided with what was becoming technologically 
possible. Electronics is characterized by an innovation curve to be 
found in no other industry: its products become cheaper and 
smaller while at the same time becoming more powerful. This 
unbeatable mix is underpinned by the speed of development in the 
technology of silicon chips, the heart of any electronics product. 
Throughout the 1970s and 1980s, it was commonplace for the 
power of a particular family of chips to double every two to three 
years. Apple could put on a desk a machine costing $1,500 in the 
early 1980s which in many respects was as powerful as a main- 
frame computer of a decade earlier. Apple seized the possibilities 
opened up by this rate of innovation to transform itself from a two- 
man partnership operating out of a garage in 1976 to a multina- 
tional corporation with sales of more than $2 billion a decade 
later. 

Britain was the first country after the United States to catch the 
home computing bug; but British manufacturers initially concen- 
trated on simple machines, less sophisticated than Apple's prod- 
ucts and mainly suitable for computer games. This market was 
supplied in North America by companies like Commodore and 
Atari. Clive Sinclair sparked the craze in Britain by launching in 
January 1980 the ZX80: priced at £99.95, it was billed as the 
world's smallest and cheapest computer. Some 20,000 were sold 
by September 1980, while a year after its launch the company was 
shipping about 9,000 a month. 

The success of each new Sinclair computer seemed to surpass 
that of its predecessor. More than 250,000 ZX81s had been sold 
by December 1981, nine months after its launch. At the start of 
1983, Clive Sinclair cashed in on his success by selling 10 per cent 
of his company, Sinclair Research, in a deal which valued the 
concern at £136 million. In December 1983, Sinclair proudly 
announced that a million Spectrums had been sold in the twenty 
months since its launch. 

Home computers were the novelty gift for Christmas 1982, and 
many high street stores sold out of their stock ten days before the 
start of the Christmas break. It seemed as though every parent of 

• 136- 



ALAN sugar: the amstrad story 

teenage children was being badgered for this latest wonder prod- 
uct, which was cleverly projected as both fun and an essential 
educational aide. The home computer fever raged more strongly 
still in Christmas 1983. Acorn boasted of 200,000 orders, while 
Sinclair Research said it could have sold 50 per cent more than 
the 300,000 machines it shipped. 'I don't remember a product 
which has developed so rapidly in such a short time. ... It was 
beyond the bounds of all expectations,' gushed Peter Jackson, 
marketing director of Rumbelows, as he surveyed the Christmas 
season. 

One might suppose that Sugar's decision to enter such a 
buoyant market would have been applauded, but that was not 
how his initiative was viewed by many people. Customers like 
Comet, Amstrad followers in the City and commentators in the 
press - all queried Sugar's judgement in attacking an extremely 
competitive market which appeared to be dominated by a clutch 
of nimble companies. Before the CPC464 was launched, Comet's 
Nick Lightowler and Sugar talked over Amstrad's plans on the 
phone: 'It was a bloody great worry for me at the time,' Lightowler 
recalls. 

The City registered its nervousness by responding badly to a 
rights issue by Amstrad in May 1984, a month after the launch of 
the CPC464. A rights issue is a means of raising fresh capital for a 
company by giving existing shareholders the right to buy new 
shares. If, as often happens, many existing shareholders forgo their 
rights, then the new shares can be sold on the open market. 
Amstrad wanted to raise £12.9 million by issuing one new share 
for every six in existence to finance the development costs of the 
CPC464 and an extension to the Shoeburyness factory. Sugar 
passed up his entitlement to buy his slice of the new shares, which 
resulted in his stake in Amstrad falling from 58.9 to 50.5 per cent. 
Even though the new shares were offered at 85p each, 17p less 
than the closing price on the day when the issue was announced, 
only 83.3 per cent of them were bought - a snub from the City for 
Amstrad. 

In the event, the City was right to be nervous about the 
immediate prospects of Amstrad's foray into computers - but for 



137 



DAVID THOMAS 

the wrong reasons. Amstrad's first taste of the computer business 
was characterized not by strong competitors but by a collapsing 
market. In the opening months of 1984 few people realized that 
Christmas 1983 had been the high watermark of the home 
computer craze in Britain. True, a cloud had appeared on the 
horizon in the summer of 1983 when an abrupt end to the personal 
computer boom in the United States triggered a round of price- 
cutting there, but the computer industry in Britain ignored this 
warning signal. The change of mood became apparent in Britain 
shortly after Amstrad launched its computer. 

A few auguries of the problems to come emerged in the run-up 
to Christmas 1984. The summer had seen the disappearance of 
two smaller computer companies, Dragon and Torch, the latter 
taken over by Acorn. In October, Acorn cut its prices and began 
to issue advertisements attacking Sinclair - a turn of events which 
led to a well-publicized exchange of blows in a Cambridge pub 
during a pre-Christmas drinks session between Sir Clive Sinclair 
and Chris Curry, Acorn's managing director. 'We party a lot in 
Cambridge,' a Cambridge high tech entrepreneur wryly confessed 
to the Wall Street Journal, while Amstrad reacted with a tongue-in- 
cheek announcement of a new computer game called 'This Busi- 
ness Is War'. The game featured a bespectacled Sinclair-character 
throwing computers at a Curry-figure. 'We're indebted to Sir Clive 
and Mr Curry for offering this wonderful chance to extend our 
software range,' Amstrad told the Daily Mail. 

Christmas 1984 turned into a disaster. Retailers had tried to 
avoid the problems of the two previous Christmases (when 
demand heavily outstripped the supply of computers) by placing 
large orders with all the main manufacturers. Most of the 200,000 
Amstrad CPC464s sold in 1984 were snapped up during the 
Christmas run-up. In the event, demand was weak amid much 
talk about the home computer market in Britain having reached 
saturation point. Manufacturers churned out 1.7 million home 
computers during 1984, but the public bought just 1.35 million. 
Retailers were left with substantial stocks after Christmas, which 
meant that orders in the opening months of 1985 dried up. 

It soon became plain that the British home computer industry 



138 



ALAN sugar: the amstrad story 

was heading for a shake-out. 'There are too many manufacturers 
with too many machines chasing too small a market, and they are 
probably not the products people want,' warned John Rowland, 
computer buyer at W. H. Smith, in February 1985. Some manu- 
facturers agreed: 'The popular games-playing market has become 
a very uncomfortable place to be. Price competition wall be 
horrific. It is not a market we want to be in for very long,' said 
Acorn's Chris Curry. 

The bad news piled up as 1985 unfolded and personal computer 
companies faced the financial headache of having to nurse huge 
stocks. Prism, a large Sinclair distributor, went into receivership, 
followed by Oric, a second-string home computer manufacturer; 
Acorn's shares were suspended while the company worked on a 
reorganization plan; Sinclair Research postponed long-standing 
plans for a flotation. These setbacks all struck the industry in the 
first two weeks of February. Within the next three months, Acorn 
was rescued by Olivetti, the Italian office products group, while 
Clive Sinclair began his long search for a capital injection which 
was to culminate in Amstrad's intervention in 1986. 

Computer companies had been the darling of the City during 
the home computer boom and their fall from grace was sudden 
and violent. Amstrad suffered by association with the troubles of 
its competitors and its share price - in the doldrums since the 
badly received rights issue - was dragged down by the computer 
industry debacle to a low of 64p. Alan Sugar tried to reassure the 
stock market: 'Some companies may have had great ideas during 
the boom, but lacked the business expertise to manage properly 
when it faded. We don't have any stocks at this time of year. In 
the summer, we'll be getting brown in our deck-chairs, not 
counting up all the computers in our warehouse,' he told the 
Financial Times in June 1985. 

The stock market's confidence in Amstrad began to recover only 
after August 1985, when Sugar unveiled his first word processor 
and the share price moved back above 85p. Faith was finally 
restored in October, when he announced results for the 1984-85 
financial year ending in June 1985. Profits jumped by 121 per cent 
to £20.2 million, on a 60 per cent increase in turnover to £136.1 



139 



DAVID THOMAS 

million. These figures were enough to convince the most hardened 
sceptic that Amstrad had indeed avoided the bloodbath among its 
competitors: the company's shares were marked up 16p to 134p 
on the day of the announcement. 

Most remarkable of all was the transformation within the 
Amstrad business which these global figures concealed. In 
1983-84 audio, colour televisions and video recorders accounted 
for 96 per cent of Amstrad's sales; the following year, these lines 
represented only a third of the company's turnover. In 1983-84, 
exports made up 13 per cent of sales; the following year, they 
accounted for 53 per cent of turnover. A new Amstrad had been 
created by its launch in 1984 - not just into computers but also 
into some key Continental markets. 

Two points had escaped the pundits who bracketed Amstrad 
with the Sinclairs and Acorns of the world. First, the greatest pain 
was felt by home computer companies at the bottom end of the 
market. The buying public had become bored with computers 
which were no more than games machines. As British consumers 
became more sophisticated about computers, so they began to 
demand machines which offered business applications like 
accounting and word processing. By giving the CPC464 one foot 
in the business market, Amstrad was ideally placed to benefit from 
this flight to quality. Sugar welcomed the greater clarity in the 
home computing market, as he made clear towards the end of 
1985: 'We've come to the end of an era. By March all these cheap 
flog-off lines will have gone. They can't be made any more at those 
prices and I will be glad to see the garbage and junk out of the 
market.' 

The second factor buoying up Amstrad was its growing overseas 
success. Sugar had been building up foreign sales of the CPC464, 
particularly in France, Spain and West Germany. The Amstrad 
chairman had made no secret of this, emphasizing to the Financial 
Times in February 1985 that the company was exporting many 
computers: The only country we are concerned about is the U.K. 
because the doom and gloom is making the dealers nervous. 
Computers are not like CB radio, a market which stopped as if 
someone had turned off a light switch.' 



140 



ALAN sugar: the amstrad story 

But the British computer industry and its camp-followers were 
insular. At that time, few of the industry's leaders like ICL, Acorn 
or Apricot had significant overseas sales. In consequence, few 
people were able to judge Sugar's claims of overseas conquests. 
The truth emerged only later: while the British market was in the 
doldrums in 1985, Sugar had been selling thousands of his 
computers on the Continent. This left him in a unique position 
later in the year to resist the demands from British retailers for 
price cuts. 

Alan Sugar gives a vivid account of this period: 

The usual array of predators, such as Dixons, W. H. Smith and 
Boots, were hovering around like the praying mantis, saying, 
'Ha, ha, you've got too many computers, haven't you? We're 
going to jump on you and steal them off you and rape you when 
you need money badly, just like Uncle Clive.' And we said, 'We 
haven't got any.' They didn't believe us, until such time as they 
had purged their stocks and finished raping Clive Sinclair and 
Acorn, and realized they had nothing left to sell. So they turned 
to us again in November of 1985 and said, 'What about a few of 
your computers at cheaper prices?' We stuck the proverbial two 
fingers in the air, and that's how we got price stability back into 
the market. They thought we were sitting on stockpiles and they 
were doing us a big favour. But we had no inventory. It had 
gone to France and Spain. 



141 



8 



Birth Pangs in Europe 



'Alan, I have to tell you something. I am waiting for a baby. And 
I must tell you the truth, it will be a very difficult job for me to get 
this baby.' 

This was not the best news that Alan Sugar had ever received. 
Amstrad's first and only employee on the Continent had tele- 
phoned him from Paris to say in her best English that she was 
pregnant. As though that were not bad enough from his viewpoint, 
the new 32-year-old managing director of Amstrad France would 
have to spend about three months in hospital because her doctors 
were predicting a difficult birth. 

Marion Vannier still remembers vividly the call she had to 
make to Sugar back in 1982. A thin, elegant woman with short red 
hair and quick movements, she darts round her office answering 
phones while telling her story. Even now she breathes hard, 
pausing between each word, as she recalls summoning up the 
nerve to give Sugar the news. 'You cannot imagine how I felt. I 
could not tell him. I couldn't. Each time I thought about telling 
him, I was crying. It was impossible to tell him.' 

Only a few months had passed since Sugar had phoned her with 
the proposal: 'Marion, how would you like to form a French 
company for me?' 

The offer could not have been better timed because Vanniers 
employer was folding up. She had spent the past ten years working 
for Cogel, a concern which distributed electrical goods in France. 



142* 



ALAN sugar: the amstrad story 

She had made the most of that decade, joining as a junior assistant 
and rising to become commercial director after three years. But in 
1980 her bosses sold out to a foreign company; the relationship 
never gelled and early in 1982 the new owners announced that 
they were liquidating their recently acquired French operation. 

So there was no reason for Vannier to hesitate when she heard 
Sugar's proposition. After all, she had been selling Amstrad's 
audio equipment in France since the early 1970s; she knew Sugar 
well and understood the French market. Exports had been import- 
ant to Amstrad throughout the 1970s. Overseas sales had 
accounted for between 10 and 25 per cent of turnover since 1973, 
after Vannier's company started pushing Amstrad's equipment in 
France. 

Yet Sugar had had to expend virtually no effort on his French 
exports, only visiting the country perhaps once a year to exhibit at 
the main French audio fair. He was able to leave the hassle and 
expense of distributing Amstrad's equipment in France to Cogel, 
a pattern he was to repeat in his early forays into Spain, West 
Germany and the U.S. That he broke this pattern as early as 1982 
in France by setting up his own subsidiary is easily explained. He 
had come to trust Vannier's business judgement and did not want 
to lose her. But clearly she lacked the capital to set up as an agent 
on her own, so he decided to establish his own subsidiary with 
Vannier as its boss. 

The Amstrad business had been relatively small beer through- 
out the 1970s for Vannier, representing just one of the many lines 
which her company distributed. She had needed to visit Britain 
only twice a year to look at Sugar's latest goods and to check out 
the trade fairs to see what other manufacturers were producing. 
Indeed, it was at one such fair in 1972 that she had first met 
Sugar, hawking an early line of Amstrad amplifiers. An agreement 
to be Amstrad's exclusive agent in France followed shortly 
afterwards. 

Wanting a breathing space from the traumas of helping to wind 
down Cogel, Vannier did not expect radical changes in her life 
when she accepted Sugar's proposal to set up an Amstrad subsidi- 
ary in France in March 1982. In June she launched Amstrad's 



143 



DAVID THOMAS 

French operation from a one-person office above a boulangerie on 
the main road through Sevres, a little town on the outskirts of 
Paris. The first headquarters of Amstrad France boasted a tele- 
phone, a telex and not much else. Alan Sugar had given her a 
cheque for 50,000 francs (about £4,500) and a free hand, expecting 
her to get on with it. By then, however, she was well and truly 
pregnant. 

Over the next year, Marion Vannier would have to draw on all 
the reserves of strength she had built up since her first marriage 
collapsed in 1971. She had enjoyed a comfortable upbringing in 
Charentes, a deeply traditional part of south-west France dotted 
with small towns and the chateaux of cognac producers. Her father 
was a director of a brandy company; her mother, it goes almost 
without saying, did not work and Marion was educated in a 
convent to expect a similar existence. True, there were signs of 
spirit in the young girl as she grew up: she had a habit of choosing 
the most headstrong horses to ride through the lush Charentes 
countryside and she was good enough at gymnastics to compete in 
the French championships. But everything seemed to be going 
according to pattern when she became an 18-year-old bride in 
1968. She had every reason to anticipate a lifetime of playing 
tennis and bridge and bringing up her children - a way of life 
shared by hundreds of thousands of other French middle-class 
women. 

In fact, she had traded the straitjacket of rural French life for a 
set of equally claustrophobic, almost feudal obligations by marry- 
ing the heir to a cognac fortune and moving into his chateau. She 
had to attend mass regularly, gloved and with a silk scarf covering 
her hair. 'I had a beautiful garden, I rode and I played tennis, yet 
I could not fill my days,' she remembers. 

Three years later, aged 21, she cut through her increasing 
boredom by taking a train to Paris. 'I left my bags with a friend of 
my father. Forty-eight hours later, I found a job as a switchboard 
operator in a bank near the Champs-Elysees. I earned 850 francs 
a month and paid 400 francs for my room. But I was free.' 

Vannier was having to earn her living with no obvious qualifi- 
cations for doing so. After a string of short-lived jobs, she answered 



144 



ALAN sugar: the amstrad story 

an advert from the electrical distributor, Cogel, with whom she 
was to spend the next ten years. It said that the company was 
looking for an assistant who spoke English. She had all but 
forgotten the smatterings of English learnt at school, but she went 
along in any case. 

As she entered Cogel's premises, she was greeted by a man who 
started speaking to her in English, and she laughs now as she 
describes her struggle to reply. 'Then he mentioned the VAT 
problem. How do you calculate the VAT on top of . . . ooh?' she 
rolls her eyes as she recalls the impossibility of the question. 

Vannier's reply was simple and to the point: 'Really, the truth 
is I don't know about your VAT thing, but if you explain it to me, 
then in three minutes' time: no problem.' And so it proved. She 
quickly made herself indispensable. After three years she replaced 
the commercial director, who had been traumatized by her greed 
for work. 

But in 1982 Vannier was well aware that the combination of 
giving birth to her third child and to Amstrad France would dwarf 
the challenges she had faced in the 1970s. She had spent three 
months in hospital delivering her second child and her doctors 
warned her to expect a repeat performance. A simple calculation 
revealed that she would be flat on her back from October, during 
the build-up to the Christmas season - the most important part of 
the year for selling Amstrad equipment. 

Vannier decided that the only answer was to move her office 
into the hospital, have a telex installed by her bed and give all her 
main customers her telephone number there. She conducted 
Amstrad's business from her hospital bed night and day. To help 
her she employed as a secretary a man who had worked at her 
previous company. One of his main jobs was to prepare cheques 
for her to sign, but he had another role which transcended the 
normal bounds of a secretary's relationship with a boss, since he 
happened to share blood groups with Vannier who needed a 
regular supply of blood. This man was coming to the hospital to 
make me sign the cheques and to give me blood,' she says. It is a 
moot point as to which vital fluid, cash or blood, was more 
important to her at the time. 



145 



DAVID THOMAS 

Her doctors had been correct in predicting a difficult delivery 
and Vannier sketches a large cross over her midriff when she 
describes the incisions needed to perform the Caesarian birth. It 
wiped her out. 'I was completely dead when I had the child. When 
Alan called me in the hospital a couple of days later to talk about 
business, I could not speak/ she recalls. 

But Vannier was back in her proper office in January 1983, 
where a large picture of her youngest daughter, Djinn, now hangs 
behind her desk. 'This is my Amstrad kid,' she says, pointing at 
the picture and smiling. 



While Marion Vannier was struggling with her new business from 
a hospital bed in Paris, Jose Luis Dominguez in Madrid was faced 
with a problem which, from a commercial point of view at least, 
was even more fundamental: which business to go into. The 
wisdom or otherwise of his decision would not only determine 
whether or not Dominguez would make his fortune. It would also 
decide whether he was ever to honour a vow made to his dead 
father. 

Like Vannier, Dominguez came from a business background; 
his father's firm supplied aluminium to construction companies. 
Also like Vannier, he had to make his own way in the world. For 
as the youngest of five brothers, he knew from an early age that he 
could not expect a senior place in the family firm; and to remove 
any lingering doubt, the family business crashed when Dominguez 
was 17. 

That prompted him to leave school and find work. He had been 
studying economics and business, more because his father thought 
it would help him to get a job than out of any natural conviction. 
But when he left without finishing his studies, the only job he 
could find was as a door-to-door salesman, touring the big housing 
estates on the outskirts of Madrid and selling anything that came 
to hand - encyclopaedias, insurance, English courses. 

Unlike Vannier, the young man was not naturally self-confident 
and had to force himself to knock on the doors. Dominguez is 
squat and bearded, an Alan Sugar lookalike, but in his case the 



146 



ALAN sugar: the amstrad story 

powerful physique is softened by a diffident half-smile playing 
permanently on his face. 'This work on the estates was very 
difficult for me, because I was a shy man.' he recalls. 

Dominguez' father urged him to drop this work in favour of 
something more secure and respectable, like a job in a bank, but 
he was driven by the suspicion that his father believed him to be 
incapable of making it on his own. The young Spaniard became 
obsessed by the need to prove himself. 'I was trying to demonstrate 
to my father that I could be successful by making money,' he 
explains. 

Slowly Dominguez saved up the commission he earned by door- 
to-door selling to the point where he reckoned he had enough to 
impress his father with his diligence. But the day before Domin- 
guez was planning to produce all his hard-won cash for his father 
to admire, he heard that he had been killed in a car accident. 'On 
that very day, I swore to myself that I would make a successful 
business for my father.' 

Dominguez had to wait ten years before he was able to carry 
out that promise, however. In his early twenties he drifted into the 
insurance business, rising in five years to become commercial 
director in Spain for Nationale-Nederlanden, the big Dutch insur- 
ance group. He was given the job of opening offices all over Spain 
for his Dutch bosses, which was invaluable experience for his later 
efforts in setting up a Spanish distribution network for Amstrad. 
But it was also plain to him that he could not expect to go much 
further with the Dutch group. 

In 1981, by then in his early thirties, Dominguez returned to his 
earlier dream of striking out on his own and set up a trading 
company which he named Indescomp. But as he recalls: T was 
wondering what kind of business to get into. I was looking for a 
business of the future, but this was very difficult for me. Then one 
day, I read in a magazine about personal computers and their 
success in America and England.' 

Dominguez devoured the latest news about the companies 
leading the personal computer revolution - Apple and Atari in the 
United States, Sinclair and Acorn in Britain. Personal computers 



147 



DAVID THOMAS 



were then almost unknown in Spain, but it was immediately 
obvious to him that this state of affairs would change. 

Despite speaking only minimal English, Dominguez travelled to 
a computer show in Britain, at that time by far the most developed 
personal computer market in Europe. Sinclair's machines, cheap 
and simple, struck him as ideal for the Spanish market. He went 
to Cambridge to try to arrange a deal with Sinclair, only to be told 
that the latter had already signed a distribution agreement for 
Spain. 

Recovering from that disappointment, his eye lit next on Acorn, 
which was notching up big sales with its B.B.C. micro. He met the 
person who handled Acorn's international business in London, 
and quickly won the exclusive franchise to sell the company's 
machines in Spain. 

But Dominguez' first foray into the personal computer business 
turned into a disaster. He found it impossible to shift the 2,000 
Acorns which he had bought, partly because he had no money left 
for adequate promotion of the machines but also because Spain 
was not quite ready for the mass introduction of personal 
computers. 

Dominguez changed tack. Let others endure the hassle and risk 
of launching the personal computer revolution in Spain; he would 
follow at a safer distance, one step behind. In future, he would 
supply the gamut of peripheral equipment that the Spanish users 
of the machines would need - software, manuals and add-ons. 

For a year or so, that meant concentrating on the Sinclair 
market. Sinclair's personal computers took off like wildfire, 
helping to establish the notion of a personal computer with the 
Spanish public. Dominguez was a pioneer in persuading big 
department stores to stock all the bits and pieces needed by 
Sinclair users. He had his triumphs with writing software too, 
including one game, Bugaboo, which topped the charts in both 
Spain and Britain. 

Indescomp grew step by step, but still Dominguez hankered after 
introducing another personal computer to Spain. That was where 
the big money was to be made and he reckoned he could displace the 
market leaders, Sinclair and Commodore. He had built up an 



148 



ALAN sugar: the amstrad story 

effective distribution network and he understood the Spanish per- 
sonal computer market as well as anyone in the business. 

In 1983 Dominguez embarked on his travels once more. He 
went to Hong Kong, to japan and to the United States in search 
of his ideal machine, but none that he saw filled him with total 
confidence and he returned to Spain empty-handed. 

Then one day early in 1984 Dominguez read in a computer 
magazine that a British company, Amstrad, was planning to 
introduce a computer with a built-in monitor and cassette, and 
with just one cable. Immediately he was struck by the elegance of 
this solution to the problems faced by the average Spanish family 
in connecting up a Sinclair machine. 'In that moment I thought, 
that is a very good idea. The man who thought that must be an 
intelligent man. I thought I'd like to meet him.' 

This was easier thought than done, however. Dominguez 
phoned Amstrad's headquarters in London and was put through 
to Bob Watkins, the technical director. When he explained his 
ambition to launch Amstrad's new computer on the Spanish 
market, Watkins was less than encouraging. Amstrad had no spare 
resources to service sales in Spain; it was fully stretched trying to 
meet demand in Britain and France. 'We are not interested in the 
Spanish market,' Watkins concluded. Thank you very much and 
goodbye.' 

The only glimmer of hope held out by Watkins was the faint 
possibility that Amstrad might be ready for Spain in two to three 
years. In the ever-evolving world of personal computers, that was 
like promising to do something in a generation's time and was no 
use to Dominguez. 

But the Spaniard kept reading advance news about the launch 
of Amstrad's first computer and he kept ringing through his 
interest to Watkins, who continued to fob him off . . . until one 
day Dominguez happened to mention that Indescomp had had a 
number one software game, not just in Spain but in Britain too. 

Watkins's tone changed. He explained that Amstrad wanted as 
many good games as possible for its new machine. He was 
prepared to see Dominguez in London, but only on condition that 



149 



DAVID THOMAS 



they talked solely about software and not about distributing the 
new computer in Spain. 

Dominguez was not the first person whose heart sank on first 
glimpsing Amstrad's headquarters in Garman Road, Tottenham. 
He had been expecting a large building, replete with the elegant 
executive suites he was later to favour for Amstrad Spain. In fact, 
he found cramped offices piled high with boxes, partly in prep- 
aration for Amstrad's move to a new base in Brentwood. Tt 
created a bad image,' Dominguez remembers. But he soon learned 
that the image of their building was the last thing on earth likely 
to bother the people at Amstrad. 

Watkins told him that Amstrad needed a game from Indescomp 
in a month flat. When Dominguez asked whether he could take a 
prototype of the new Amstrad computer (the CPC464) to Madrid 
to work on, Watkins was reluctant but eventually agreed on the 
understanding that Dominguez returned in a month with a game. 
So the Spaniard gingerly carried the prototype back to Madrid. 

Dominguez put his employees to work on the software round 
the clock for a month; instead of writing just one game for the 
CPC464, they came up with two including a version of Bugaboo 
tailored to the Amstrad. At the end of the month he phoned Bob 
Watkins to give him the good news. Watkins told him to send the 
games to England by courier. 'I then took the most important 
decision in my professional life,' Dominguez recalls. 'I said to Bob, 
"No, I want to bring them to you myself" 

When Dominguez reached Amstrad twenty-four hours later, he 
had one simple request: 'I would like Alan Sugar to come and see 
the demonstration of my games. 5 

Watkins hesitated for a second before picking up the phone and 
urging Sugar to take a look at Dominguez' handiwork. When 
Sugar bustled in, shirt-sleeves rolled up and not in the best of 
humour, his manner betrayed his feelings to a fault: 'Who is this 
man wasting my time?' it said. 

This less than fulsome welcome from a man Dominguez had 
been trying to see for three months reinforced his nervousness. 
To his horror, his games did not work when they were first 
loaded into the computer. Sugar stood there scratching his beard. 



[50- 



ALAN sugar: the amstrad story 

Eventually, however, the demonstration went smoothly and 
Sugar seemed impressed. 'O.K. we'll fix up a royalty deal with 
you.' 

Dominguez then made another bold move in his battle for 
Amstrad acceptance: 'I don't want to sell the games to you. You 
can have them as a present.' 

Nothing in Sugar's business experience had prepared him for 
presents and his surprise shone through. 'I wanted Amstrad to see 
that it is possible to do good work in Spain. I know you're not 
interested in Spain at the moment, but I wanted to make the 
point,' Dominguez explained. 

Sugar immediately softened his attitude and started asking 
Dominguez about his ideas for launching the Amstrad computer 
on the Spanish market. Negotiations started there and then over 
Dominguez becoming Amstrad's agent in Spain. 



Marion Vannier was building a respectable business for Amstrad 
in France. She had been able to sell about 20,000 Tower Systems 
- then Amstrad's flagship audio product - in 1982-83. But it was 
very much a small-scale enterprise: Amstrad France still employed 
only two people, Vannier and her secretary. 

Sometime late in 1983, Sugar had explained to Vannier about 
his plans for moving into the computer market. She put this to the 
back of her mind, assuming that she would handle the business 
just as she was responsible for Amstrad's audio lines. 

One day early in 1984, however, she took a call from Sugar: 
'Marion, I have to speak frankly. You are not well enough 
established to handle this micro business. It's much more compli- 
cated than our audio stuff. You will have to translate all the 
manuals; you will have to have technical people. You are not 
ready for it.' 

Vannier felt confident enough to handle anything. But she did 
not argue. 'Alan is a difficult man to argue with,' she says. 

Sugar was back on the phone a few days later. Tve been 
contacted by a French guy who is handling Oric (another com- 
puter manufacturer) in France. He's ready to be our agent; he's 



■ 151 



DAVID THOMAS 

got forty people already. It's really a better idea for you to handle 
the audio business and for him to handle the micro business.' 

Vannier did not take to this proposal at all and decided to fight 
back. But all she said to Sugar on the phone was that she would 
come over to Amstrad headquarters with the Oric man to talk 
things over. In the late spring of 1984, Vannier and her new rival 
duly flew to London. 

Sugar told his prospective French agent that he would not get 
any of the new machines until the end of December, because 
Amstrad was fully booked with British orders. This did not please 
the Oric man, who insisted he needed some machines for the 
Christmas season. After much haggling, Sugar gave him 4,000 
CPC464s for October and November. 

Back in her office in Sevres, Vannier phoned Sugar: 'Why do 
you want to give him an exclusive deal? Give me the chance to 
have the same number of units as him. You gave him 4,000, so 
give me 4,000 as well. You'll soon see who is the best.' 

Sugar saw the sense in this. After all, Vannier was an Amstrad 
employee - why not see what she could do? When he phoned the 
Oric man with an alternative proposal for him to handle the 
CPC464 in France on a non-exclusive basis, the man objected 
strongly. But Sugar had become convinced that Vannier should 
handle at least part of the business. However, neither Vannier nor 
her rival from Oric would give way and in the end the Oric man 
disappeared from the Amstrad scene. 

Sugar decided to give Vannier 8,000 machines - her own quota 
plus that of her erstwhile rival. In June he phoned her with the 
news. Tm sorry, Marion. It's really only a token quantity, but 
that's all I can spare.' 

Vannier was secretly pleased she was not being allocated any 
more machines. Her baby was still small and she was looking 
forward to a summer holiday without worry, so she resolved to sell 
her batch before going away. In any case, 8,000 personal com- 
puters seemed enough to her; she knew nothing about them - not 
even what a micro-computer looked like - until her consignment 
of CPC464s arrived from Britain. 

Vannier took a sample of the new machines around to show her 



152 



ALAN sugar: the amstrad story 

best customers, the seventeen big department stores who bought 
the bulk of her audio equipment. But she was in for a shock. None 
of them had yet moved into the personal computer market. They 
all ordered a token quantity, five at most, more to maintain good 
relations with her than because they really wanted these novel 
machines. 'Seventeen multiplied by five. You can imagine how I 
felt. It was a horrible disaster,' Vannier says. 

So she did not spend a worry-free vacation after all. She had 
quickly realized that for the CPC464 she would need to employ a 
completely different form of distribution from that which she had 
perfected for Amstrad's audio equipment. Most of the personal 
computers sold in France at that time were channelled through 
small specialist retailers but Vannier - still virtually a one-woman 
band - could not conceivably trek round these myriad outlets 
herself 

Once back from holiday, she advertised the CPC464 in the 
French computer press, listing her seventeen main outlets as 
places where the machines could be bought. But, worried that the 
seventeen were concentrated in Paris, she also used the advert to 
urge the public to ask any computer retailer for the machine - a 
bold stroke considering that no computer retailer in France then 
stocked the CPC464. 

The French public followed the British in liking the CPC464. 
Taking her cue from the way in which the machine had been 
promoted in Britain, she advertised it in France as 'The Anti- 
Crisis Home Computer', stressing the convenience of not having 
to unplug the computer so that the rest of the family could watch 
television. Free from the need to charge a distributor's mark-up — 
typically around 30 per cent - the Amstrad machine sold in France 
for the same price as in Britain. 

Computer shops all over France received orders for the CPC464. 
Never having heard of Amstrad, the retailers offered customers 
Sinclair, Oric, Commodore, virtually anything else instead. But 
the customers stuck to their guns and orders flowed into Sevres 
from small shops all over France. Amstrad U.K. even managed to 
find its French subsidiary a few thousand extra machines to help 
satisfy Vannier's pre-Christmas rush. 



153 



DAVID THOMAS 

The Frenchwoman did not know whether to laugh or cry. She 
had been used to handling seventeen large customers, each of 
whom would take a van-load of goods. Now suddenly she had to 
handle hundreds of new customers, most of whom ordered in 
batches of half a dozen. And Amstrad France still consisted of 
three people: Vannier, her secretary and a new technical specialist. 
'I had no reps, nobody on the road,' she says. 

Marion Vannier remembers the next year as the worst in her 
life. Day after day, she struggled with the flood of orders with the 
minimum of office back-up. Night after night, she rushed home, 
cooked her children dinner and then returned to the office when 
they were tucked up in bed. Her second husband, an insurance 
executive, confesses: 'We decided to spend our dinners together, 
for otherwise we would no longer have seen each other.' 

By March 1985 Amstrad France had sold 30,000 computers - 
the majority during the months after Christmas, traditionally a 
quiet time for home computers like the CPC464. The company 
grew fast, expanding to twenty-five employees by June 1985. But 
even this recruitment was at first just another headache for 
Vannier who was having to do everything: hire new people, find 
larger offices and supervise the sale and distribution of Amstrad's 
machines. She was running so fast that she had no time to stop 
and admire the extraordinary explosion of sales over which she 
was presiding. 

In the year to June 1984, Amstrad's sales in France totalled 35 
million francs (about £3 million). In the following year, thanks to 
the CPC464, they rang in at 291 million francs (about £24.3 
million) - an eightfold increase in twelve months. 



Jose Luis Dominguez learned the hard way that Alan Sugar strikes 
a tough deal. The Spaniard had to agree to two conditions before 
he was appointed Amstrad's agent in his home territory. 

Sugar was so taken with the games given him by Dominguez 
that he told the Spaniard, Til give you exclusive distribution of 
Amstrad in Spain, but in return you must give me ten games.' 

That was condition number one, and it turned out to be the 



154- 



ALAN sugar: the amstrad story 

easy part of the bargain. Dominguez agreed, but remembers with 
a smile that those ten games took two years to deliver. Every 
month Sugar would phone up with the same question: 'What's 
happened to the new game?' 

Condition number two was the difficult part. Sugar refused to 
give Dominguez any credit; the Spaniard had to pay in advance 
for all the machines he took, which meant opening a letter of credit 
four months before he saw the machines, allowing for delivery 
from Hong Kong. 

Dominguez wanted to take about 500 CPC464s initially. But 
the value of his house at the time was enough to guarantee only 
about a quarter of the money he needed to finance the purchase, 
so he trawled all the banks in Madrid for credit to bridge the gap. 
In the end he put together a credit package from twelve different 
banks, but not before he started receiving nervous phone calls 
from Sugar demanding to know what had happened to the letter 
of credit. 

The machines arrived in time for Christmas 1984 and Domin- 
guez placed some adverts in the computer press, promoting the 
CPC464 as the next step up from the Sinclair and the Commodore. 
He would have liked to have gone further but, as he explains, 'I 
had no money to launch a big advertising campaign.' 

Even so, the adverts were a success and Dominguez could meet 
only about a third of the demand that Christmas. Sugar's attitude 
to the Spanish market began to change. The following summer he 
invited Dominguez to dine with him while he was holidaying at 
Los Monteros in Marbella. Over the meal, the pair discussed 
prospects for the 1985 Christmas season in Spain. 

'The credit arrangements are a big limitation on growth in 
Spain,' Dominguez told Sugar. 'I would like to take 50,000 
machines for this Christmas, but I can pay for only 2,000 with my 
own money.' 

Sugar took some convincing, still not sure that Spain had that 
much potential. But in the end he agreed to alter the credit 
arrangements and gave Dominguez 90 days to pay for any 
purchases. 

In the early autumn of 1985 Dominguez duly took delivery of 



155 



DAVID THOMAS 

his 50,000 machines, a mixture of the CPC464, CPC6128 and the 
new PCW8256 word processor. By Christmas they had all been 
sold. Amstrad's Spanish sales jumped from £4.4 million in the 
financial year 1984-85 to £32.3 million in 1985-86. Amstrad had 
arrived in Spain in sensational style. 



Amstrad was in on the beginning of the first wave of mass personal 
computing in France and Spain, just as it had been in Britain. On 
the back of the huge following for the CPC464, the company 
rapidly became the market leader in volume sales of home personal 
computers in both countries. The CPC464 sold well in West 
Germany too, notching up sales of £32.5 million in 1984-85; but 
it was largely thanks to the efforts of Marion Vannier and Jose 
Luis Dominguez that Amstrad's overseas sales soared from just 
over £10 million in 1983-84 to £72.8 million in 1984-85 and 
£174.9 million in 1985-86 - a stunning achievement. 

France and Spain were by no means identical markets when 
Amstrad launched the CPC464. In Spain, where personal comput- 
ing barely existed, the price of the machines was a predominant 
consideration. By contrast, France had a thriving hobbyist sector 
populated by techno-freaks who were happy to expend money and 
time on procuring computer, cassette, plug, screen and add-ons 
and putting them together. 

What both France and Spain lacked was a mass personal 
computing market in which hundreds of thousands of individuals 
with no interest in the interior of the machines could buy a 
personal computer from their local department store, take it home 
and make it work with the same minimum of fuss they expended 
on a new television or hi-fi. Amstrad was at the forefront in 
creating that market in both countries. 

Marion Vannier and Jose Luis Dominguez had contrasting 
styles in their day-to-day conduct of business - a difference which 
helps to explain why, in 1989, it was Dominguez who parted 
company with Amstrad. After Amstrad France took off, Vannier 
moved into a rigorously functional office block overlooking the 
single room from which she launched Amstrad France. Sevres, her 



156 



ALAN sugar: the amstrad story 

base, is a bustling little town whose half-completed office blocks 
and busy streets are a testament to the prosperity it is drawing 
from Paris - a workaday setting for Amstrad France, much like 
that of Amstrad's U.K. headquarters in Brentwood. 

When Vannier comes out of her nondescript office to greet her 
visitors, it is not immediately obvious whether this young woman 
is the boss or a secretary. Her clothes are elegant but simple - a 
plain blouse and short skirt. There is no hint of power dressing to 
impress visitors, either with the wealth she has amassed thanks to 
her success at Amstrad or with her position as one of the best- 
known businesswomen in France. 

By contrast no one could mistake Dominguez' standing in 
society, not least because he is followed wherever he goes by a car 
full of young men in dark glasses with padded suits. In Spain these 
armed bodyguards are essential to shield the wealthy and famous 
from kidnapping or terrorist attack. Yet there is a hint in Domin- 
guez' manner that he regards this need for permanent protection 
as final proof of his rise into the ranks of Spain's super-rich. 

Under Dominguez, Amstrad Spain's headquarters were in a 
fashionable quarter of northern Madrid, not far from the royal 
palace. There was no chance of mistaking anyone but Dominguez 
for the boss as he sat in this building in the summer of 1989, shortly 
before he was effectively forced out of Amstrad. He had a connecting 
suite of rooms made up of exhibition room, meeting room and a 
sitting room with a fine view towards the centre of Madrid. Domin- 
guez sat at the far end of this suite in splendid isolation behind a 
large desk. The Spaniard, aware that this extravagance ran against 
the Amstrad corporate culture, joked that his offices were more 
lavish than the company's headquarters in Britain. 

Dominguez' success in building up Amstrad Spain made him a 
star of the Spanish business scene. He was frequently profiled as a 
leading representative of the more aggressive entrepreneurial spirit 
which has pervaded the Spanish economy in the 1980s. His views 
on marketing, a key obsession of modern Spanish business, were 
particularly canvassed. 

Vannier too has been feted as one of the very few successful 



157 



DAVID THOMAS 

female entrepreneurs in France. Named as French businesswoman 
of the year in 1988, her rise from 21 -year-old divorcee with no 
qualifications to business tycoon has regularly been depicted as a 
triumph of mythical proportions. With its penchant for English 
phrases when writing about business, the French press describes 
Vannier as 'un success story' and 'une star du business'. Portrayed 
as the symbol of a new wave of female managers in France, she is 
consulted on any issue linking business and women. In the space 
of a few weeks in 1989, for example, she turned down not only a 
request by one of France's biggest advertising agencies to star in a 
high profile promotional campaign for Le Monde newspaper, but 
also an invitation from Simone Veil's centre party to run as a 
candidate in the European elections. 

Both Vannier and Dominguez were naturally pleased with the 
fame that success brought them in their native countries. Their 
work for Amstrad also made them wealthy in the late 1980s. 
Vannier's fortune came from generous share options, while Dom- 
inguez sold his distribution company, Indescomp, to Amstrad 
when Sugar decided to create subsidiaries in all Amstrad's main 
spheres of operation. But Vannier gave the impression to visitors 
of wearing this fame more lightly. There was no suggestion of false 
modesty in her explanation for her high media profile: Journalists 
here are badly in need of a successful businesswoman.' 

Yet these very different personalities gave a common account of 
the reasons for Amstrad's initial success in the mid-1980s in both 
countries: total attention to marketing and sales was at the heart 
of their stories. Both were decisive in persuading the large chains 
and department stores - groups like Euromarche and Darty in 
France and Corte Ingles in Spain - to start selling personal 
computers, partly displacing the specialist outlets in the process. 

T am not a technical person. The technical side is too difficult 
for me,' explained Dominguez, adding proudly, T am a salesman.' 
His selling instincts helped to make the name Amstrad almost 
synonymous with low end personal computers in Spain. 

Vannier had to delegate responsibility for many of the tasks she 
was used to doing herself as Amstrad France grew to 125 people 
by 1989. But marketing and publicity were the two aspects of the 



158 



ALAN sugar: the amstrad story 

Amstrad business she retained, spending up to 5 per cent of her 
turnover on advertising. 'To talk straight to the final consumer. 
That is 90 per cent of my sales effort/ she says. 

The marketing approach favoured by both Vannier and Dom- 
inguez during the mid-1980s was basically the same as that honed 
in Britain by Alan Sugar. Vannier was able to translate advertising 
slogans like 'compatible avec qui-vous-savez' (compatible with 
you-know-who) - used to sell Amstrad's first I. B.M. -compatible 
personal computer, the PC 151 2 - straight from English into 
French. Her description of Amstrad France could have been taken 
from the mouth of Alan Sugar: 'We transform elitist products into 
mass products. We seed new markets.' 

The Amstrad emphasis on simple-to-use products, powerfully 
promoted and priced at a level which most consumers could 
afford, is a business concept which has travelled well across 
European boundaries. As Alan Sugar surveyed the take-off of 
Amstrad sales in France and Spain in the mid-1980s, he was able 
to reflect that a philosophy forged in the tough world of the low 
end British audio market was as effective in Paris and Madrid as 
in London. The question facing Amstrad as it entered the second 
half of the 1980s was this: would the company be able to build on 
its international success in France and Spain to forge a genuinely 
multinational operation? 



159 



• V>| • 



The Amstrad Effect 



Alan Sugar felt restless as the plane cruised high above Japan en 
route for Hong Kong. It was July 1984 and only three months 
since Amstrad had launched its first computer, the CPC464, yet 
the glimmer of an idea for a new line of computers had been 
gnawing away at him while he toured his suppliers on one of his 
regular Far East trips. 

He turned to Bob Watkins, sitting next to him. 'Have you got 
any paper?' he asked. Amstrad's technical director shook his head. 
'Well, what's that you're looking at? Give it to me for a second,' 
said Sugar. 

Watkins handed his boss the document he had been reading, a 
contract between Amstrad and Digital Research. Sugar took out a 
pen, turned over the paper and sketched away busily for a couple 
of minutes. 'What do you think of that?' he asked, showing 
Watkins his handiwork. 

Sugar had drawn a computer which took the all-in-one philos- 
ophy of the Tower System and the CPC464 to its logical con- 
clusion. He envisaged one box containing monitor, disk drive and 
printer. The monitor was shaped like an A4 piece of paper, the 
disk drive was underneath the monitor, while the printer was 
moulded into the machine on top of the monitor. It was this latter 
feature - building the printer into the computer itself - that was 
startlingly novel. 

The Amstrad chairman explained his idea to Watkins. A printer 



• 160 



ALAN sugar: the amstrad story 

normally has its own power supply and its own plug for connection 
to a computer. Well, why not make the printer part of the 
computer? That way, it could be driven directly from the com- 
puter, which would cut out the need for a power supply in the 
printer, remove worries about connections between printer and 
computer (the 'interface', in computer jargon) and deliver a one- 
plug machine for users. 

Watkins grasped the potential for cost-cutting, but what would 
the machine do? The two of them spent the rest of the journey 
kicking around two main options. The first was to make the 
machine a dedicated word processor. That was why Sugar had 
drawn a screen turned on its side in an A4 shape: it would make 
the screen as much like a sheet of paper as possible. The second 
option was to present the machine as an exceptionally compact 
and cost-effective computer. 

The debate flowed back and forth between the two men. 
Marketing the machine as a computer had advantages. A com- 
puter sounded multi-purpose, while a machine sold as a word 
processor might seem to have limited use. Yet the more they 
talked, the more they swung towards the notion of a dedicated 
word processor. If Amstrad could make the machine cheaply 
enough - to retail at about £400 - it could attack the huge 
typewriter market. The spread of word processors in Britain's 
offices had been slow, largely because in the recent past they had 
cost as much as £10,000. At that price, only the most well-heeled 
corporations could afford to replace electric typewriters with word 
processors. 

The Amstrad duo also foresaw marketing advantages in pre- 
senting the machine as a word processor. Many people were still 
nervous of buying a 'computer': they were unsure whether they 
would be able to understand it and what they would use it for. No 
such techno-fear would be aroused by a word processor — a 
souped-up typewriter. Indeed, many existing owners of personal 
computers used them for little more than word processing. 

By the end of the five-hour flight, Sugar and Watkins had agreed 
not only the basic design of their new machine but also how it 



161 



DAVID THOMAS 

would be projected. Amstrad was set to build Europe's first low- 
cost word processor. Not one to hang about, on 20 July 1984 Sugar 
faxed MEJ Electronics, the small team which had designed 
Amstrad's first computer, from the company's Hong Kong head- 
quarters. The fax outlined Sugar's plans for the word processor 
and attached a sketch of the machine. 

'Sorry about spelling, Regards, Alan' was the parting message 
scrawled in Sugar's hand across the bottom of the fax, which was 
headed 'Confidential: Re New Projects for 1985.' More pertinent 
than his spelling was Sugar's ability to set down on one side of a 
piece of paper the essential features of his plans for the word 
processor. 

Sugar kicked off the fax by telling his designers: 'We are 
planning to make a promotional priced word processor . . . based 
on the Amstrad philosophy of all in one unit.' He explained how 
the price would be kept low: 

We have arranged with Orion that will make FDD drive (floppy 
disk drive) in 1985 and we will not buy from Matsushita. The 
effect of this will mean we can buy a drive about half the price orig 
quoted by Hitachi. On the basis of making a major breakthru in 
price this will open a whole new world for us for the mass market. 
Orion also have big contacts with printer mechanism makers and 
they will find us a very cheap method of printer mech. With the 
facility of cheap FDD and printer mech we can go forward with 
these new projects to mass market appeal. 

Sugar then listed seven points which he wanted his designers to 
consider, adding that he was doing so 'in my very non-technical 
way' - a self-denigrating form of wording which at that time was 
a Sugar catchphrase: 

A) We dont need any sound or colour. 

B) Consolidate electronics for interfacing printer with com- 
putor. This means maybe in our existing CPC464 there is some 
additional electronics needed to make it Centronics (one of the 
most common types of interface between a computer and a 



* 162 



ALAN sugar: the amstrad story 

printer) and in the Centronics printer there is some electronics 
also. If we can delete this additional electronics we can save 
money. 

C) It is obvious that only one power supply is needed for 
monitor fdd computor and printer. 

D) Although the drawing shows the unit with 2 fdd we only 
plan 1 . The reason is that the actual word processor programme 
will be in ROM (read only memory) on the main pcb (printed 
circuit board) and as the unit is only a word processor there will be 
no need for machine operating ROM like we have in the CPC. 

E) On the basis of this unit being a complete intergrated unit 
you must disregard all industry standards of interface ie Centron- 
ics etc. This is very important to clear your mind of all these 
things. It is obvious that as this unit is a complete item there is 
nothing more to plug on to it. 

F) Consider custom ics (integrated circuits) and ula (uncom- 
mitted logic arrays or gate arrays) where ever possible. 

G) Crt (cathode-ray tube, which illuminates the screen - short- 
hand for the screen itself). As you can see from our drawing the 
crt is shown vertically. This is an obvious good idea if the item 
is to be used as a wp as this configuration is most like a sheet of 
paper. However as you know we use a TV grade tube with 
green phosphor. As 80 coloum mode is needed we must consider 
what kind of tube we need. We cant go for a full high res(olution) 
tube as they are too expensive. But we can ask maker to make 
half-way compromise to keep price down. 

The Amstrad chairman asked his designers to start thinking about 
the word processor so that they could discuss it as soon as he 
returned from the Far East. Or as he put it in the fax: 'I sent you 
this message so that you can ball park around in your mind what 
we will discuss on 15 Aug.' 

In the event, the precise design which Sugar had sketched on 
the plane was dropped and the A4 screen was buried in favour of 
a more conventional shape. It was discovered that moulding the 
printer into the top of the monitor would have caused the machine 
to overheat. Some of the assumptions underlying Sugar's fax were 



• 163 



DAVID THOMAS 

also disproved in the longer term: for instance, supplying add-ons 
to run on the Amstrad word processors has kept a host of little 
software companies in gainful employment. 

Yet Amstrad's word processor remained true to the core vision 
which excited Sugar high above the East China Sea - that it was 
possible to run the monitor and printer off a common set of 
electronics. Mej and Locomotive Software were ideally placed to 
flesh out this core vision, not only because of their role in the 
CPC464 but also because they had had plenty of experience of 
designing word processors (albeit more high cost machines) in 
their previous work for Data Recall. But it needed Sugar's clarity 
to prompt these engineers to design a word processor cheap 
enough to be within the financial reach of almost every small 
business or self-employed professional in Britain. It was this clarity 
which shone through his Hong Kong fax, soon dubbed by the 
Dorking engineers 'the tablets of stone'. 



Planning the word processor was one of several ways in which 
Amstrad was drawn more deeply into the computer business 
during 1984. Back in February, before the launch of the CPC464, 
Alan Sugar created a new Amstrad division, Amsoft, whose job 
was to smooth the company's entry into the brave new world of 
computers. For some time Amsoft lived in an uneasy symbiosis 
with its parent company, reflecting the adjustment Amstrad 
needed to make as it became enmeshed in computer culture. 

From the start of the CPC464 project, the differences between 
the computer and audio industries had been impressed on Sugar 
by Roland Perry and William Poel, the two Ambit employees who 
acted as project managers for Amstrad's first computer. Com- 
panies like Sinclair Research and Acorn had been successful 
initially because an infrastructure of services had grown up to 
support users of their machines. Since there would be nothing 
similar for the CPC464 at first, Amstrad would need to create this 
infrastructure itself. This meant writing games software and 
instruction manuals, dealing with the computer press, starting a 
users' club and publishing an Amstrad users' magazine. 



164- 



ALAN sugar: the amstrad story 

Sugar could see the sense in this, but he also realized how foreign 
it was to Amstrad's traditions. Amstrad had prospered by shifting 
boxes to the truck-driver-and-his-wife and then forgetting about 
them. It seemed poised to enter an era of shipping computers to the 
journalist-and-his-or-her-partner and then being nice to them. 

Hesitating about embracing this cultural shift, Sugar's first 
thought was to set up the Ambit people in an independent 
company with an arm's-length relationship to Amstrad. He envis- 
aged the computer boffins working from a rural retreat somewhere 
off the Mil motorway in Silicon Fen. But he soon realized that 
this would mean surrendering too much control over their oper- 
ations and decided instead to make Amsoft a division of Amstrad. 
Perry, Poel and a handful of other Ambit people left Ambit to 
form Amsoft in February 1984. 

Amsoft wanted to be separate from the rest of Amstrad, a 
geographical distance which spoke volumes about the cultural gap 
between the two groups of workers. The Amsoft people refused to 
move into Amstrad's Tottenham headquarters in the wastelands 
of North London and Sugar did not try to persuade them. Looking 
back, Roland Perry believes that Amstrad was half expecting 
Amsoft to fail: 'They wanted us to be separate from Amstrad, 
because they were not convinced it was going to be a success. If 
the whole thing collapsed in a great pile of anarchy and destruction 
like most computer companies seemed to, they wanted to be able 
to say, "Oh, that's a load of long-haired weirdos called Amsoft. 
It's not really anything to do with us.'" 

Perry and Poel rented an office from Ambit while they looked 
for a permanent home. Eventually they discovered a nondescript 
nine-storey office block 100 yards from Brentwood railway station, 
which was incongruously tacked on to a pub called the Railway 
Tavern. Its owners, the credit organization run by the Ford motor 
company, had been trying to lease or sell the building. This had 
proved an uphill task since the interior looked as though it had 
survived a nuclear attack and birds flew nonchalantly in and out 
of the upended filing cabinets and other debris. But Perry and 
Poel agreed to take the one habitable floor and told Amstrad about 
their new home. 



165 



DAVID THOMAS 

Sugar pricked up his ears, any lingering desire to keep his 
distance from Amsoft outweighed by the fact that Amstrad had 
once more outgrown its premises. He told Amsoft that he would 
look over this office block in Brentwood, a town in Essex on the 
north-east fringes of London. One of Amsoft's staff compares what 
happened next with a scene from The Godfather. A line of cars drew 
up with Alan Sugar in the lead vehicle. He and Amstrad's directors 
emerged from the cars, all wearing long overcoats; they followed 
Sugar into the building in single file, up the stairs, round two or 
three floors and out into the pub next door. Sugar sat in the pub, 
parcelling out the rooms to the directors; then he returned to 
Amsoft's offices, picked up the phone and called Ford. After a 
short discussion, he said, 'I don't care about your bloody lawyers. 
I want the keys and I want them now. If you don't have the keys 
round by four o'clock this afternoon, the deal's off.' The keys 
arrived. 

Amstrad sold its Tottenham site, moved the warehousing to its 
Shoeburyness plant and its offices into the Brentwood building, 
which remained the company's headquarters into the 1990s. 
Amstrad and Amsoft were after all to be closeted together, but 
Amstrad insisted on Amsoft having its own floor, receptionist, 
switchboard and phone number. Amstrad would be responsible 
for making, marketing and distributing the computers, as well as 
dealing with retail customers; but it would have nothing to do 
with individual computer users. Only Amsoft's phone number 
would be published, because it was Amsoft's job to field calls from 
end users. 

The Amsoft people busied themselves with user support activi- 
ties. They wrote a leaflet to accompany each computer with a 
cheery photograph of Alan Sugar welcoming users to the Amstrad 
family. 'Now you have bought the computer, why don't you join 
the club?' the leaflet gushed, introducing the company's latest 
customers to the Amstrad Users' Club. In the early days, William 
Poel even took computer journalists out to lunch. Spending time 
and money on the trade press was an activity as far removed from 
the traditional Amstrad culture as could be imagined: Trying to 
get lunch receipts paid back from the Amstrad exchequer was 



166 



ALAN sugar: the amstrad story 

regarded as almost the most daring thing that anybody in the 
company had ever tried to do,' recalls Poel, who cannot remember 
if he succeeded. 

Most of Amsoft's activities never took firm root within Amstrad. 
It soon stopped publishing games software, an activity in which it 
never distinguished itself. It sold off the users' club and the users' 
magazine, thus distancing itself from end users. The day came 
when Alan Sugar decided that Amsoft's workers ought to have 
Amstrad extensions on their desks. He had tired of having to 
resort to an outside British Telecom line in order to talk to a group 
of employees two floors below him. Eventually, the separate 
Amsoft switchboard was disbanded. 

When Poel quit Amstrad late in 1985 to run his own software 
business, he left with mixed feelings. He believes that Amstrad's 
chairman understood Amsoft's strategy, but that Sugar was him- 
self too embedded in the Amstrad culture to admit it. Poel 
explains: 

Mr S didn't like to admit to the Amstrad neanderthals around 
him that he understood what we were trying to do, because it 
was not quantifiable in pounds, shillings and pence. It was 
almost like an ethereal concept, the like of which Amstrad had 
never been party to, which said you can't pin a value on what 
these people are doing, but it's quite important to the success of 
the computer. Taking Jack Schofield (computer editor of the 
Guardian) to lunch was something which was alien to the culture, 
but I think Mr S understood it. . . . He was astute enough to 
understand these changes. Those around him took a lot longer 
to assimilate the importance of what was going on. 

Amstrad once more gave its undivided attention to what it knew 
best - ensuring that it was impossible to walk into a store owned 
by Dixons, Rumbelows and a clutch of other high street chains 
without tripping over dozens of low r -cost Amstrad computers. 
Shorn of most of Amsoft's activities, Roland Perry reverted to his 
original brief; he acted as project manager for the development of 



167 



DAVID THOMAS 

new Amstrad computers, the day-to-day link between Amstrad 
and its external design consultants. 

Besides its word processor, in late 1984 and early 1985 Amstrad 
was also working on two new computers within the CPC family. 
The first, launched as the CPC664 in April 1985, incorporated 
into the basic design the disk drive unit which came as an optional 
add-on for the CPC464. In view of its subsequent history, it is 
unfortunate that inside Amstrad the 664 was codenamed the Idiot 
(Includes Disk Instead of Tape). The idea was logical enough, but 
the project was bungled for two reasons. First, the opportunity to 
provide the 664 with additional memory was missed: 64k of 
memory was not enough to run serious business programs, the 
idea behind building in the disk drive. Second, and more import- 
antly, the market for the 664 was destroyed by the announcement 
in August 1985 of a new and clearly superior CPC computer, the 
6128. Codenamed the Big Idiot within Amstrad, the 6128 had 
128k of memory, twice that of the 664. It was also cheaper and 
more elegantly designed. 

Customers who had bought the 664 - Amstrad claimed later to 
have sold 70,000 of the short-lived machine - and dealers stuck 
with 664 stocks were naturally aggrieved at having been sold a 
turkey. In retrospect, Amstrad offers a range of reasons for its odd 
decision to rush out the 6128 just four months after the 664. 

Amstrad was under pressure from its agent in the United States 
to bring out a computer with 128k of memory as soon as possible. 
A machine with only 64k of memory would not be taken seriously 
as a computer capable of running business applications in the 
more sophisticated American market. In the event, the CPC6128 
did not sell particularly well in North America, although it became 
a flagship Amstrad product in Britain and the rest of Europe. The 
price of memory chips was falling sharply at the time, so Amstrad 
was able to launch an inexpensive 128k machine more rapidly 
than it expected. Also, some people within the company were 
affronted by the cumbersome, inelegant look of the 664, as 
technical director Bob Watkins recalls: 

I just really didn't like it, so I decided I would get our artist to 
do a rendering of something much smaller. Before the 664 was 



168 



ALAN sugar: the amstrad story 

actually launched, I was designing this other machine on the 
side. I had a sample of it made up and kept it lying around, so 
that Alan would understand the error of his ways in being so 
enthusiastic about the 664. I was careful to make the new 
machine look a mug's eyeful, nicer and more compact than the 
664. 

The 6128 was well received. A two-page test in Personal Computer 
World described a micro with 128k of memory, built-in disk drive, 
the latest version of CP/M and priced from £299 as 'outstanding 
value for money. ... As a serious home/small business machine 
it's great.' The consumer magazine Which? named the 6128 its 
'best buy' in a feature on home computers as late as November 
1987. 

But the 664 episode cast a long shadow over the 6128, winning 
Amstrad few friends. Amstrad would have to wait until after the 
launch of its new word processor to see its standing in both the 
business and computer worlds transformed. 



Chris Hall of Locomotive Software and Mej were summoned to 
the meeting foreshadowed in the 'tablets of stone' fax at Brentwood 
in August 1984 to agree the final design plans for the word 
processor. The machine was codenamed Joyce, after Alan Sugar's 
secretary of the time. William Poel was later to remark that this 
was only fitting, since Amstrad's word processor was designed to 
make Joyce and her kind redundant by inducing executives to type 
their own memos. The company's first word processor was later to 
acquire the more formal title of PCW8256 - P(ersonal) 
C(omputer) W(ordprocessor) with 256k of memory. 

Amstrad added little to Sugar's 'tablets of stone' fax other than 
to say that, like the CPC series, Joyce would use the Z80 processor 
and 3" disks. Other options were too expensive: for instance, 
Amstrad could buy the Z80 for £1 each, whereas the main 
alternative - a microprocessor from American chip manufacturer 
Intel - would have cost £10 each at the time. Since the machine 
would have CP/M as an operating system, it would also be able to 



169 



DAVID THOMAS 

run standard business software: but the marketing plan was to 
push it as a word processor and not a computer. 

Designers often wring their hands in frustration at the start of a 
major project, while their client agonizes over the precise specifi- 
cation. The roles were reversed in the case of Joyce, as Mej 
explains: 'At the first meeting, Amstrad was actually pressurizing 
us because we couldn't produce a quotation fast enough. They 
wanted us to tell them there and then how much it would cost and 
how long it would take, so that they could give us a contract on 
the spot to carry out the project.' 

The meeting was typically Amstrad - short, sharp and to the 
point. Looking back, Mej singles out the speed with which 
Amstrad launched new projects as crucial to the success of the 
word processor and other products of the mid-1980s: 'One of the 
things about these projects was how quickly Amstrad made the 
decision to do them. You would expect any other client who came 
up with an idea like the word processor still to be talking about it 
a year later.' 

Alan Sugar's commitment to speedy decision-making had been 
impressed on Mej by an incident earlier in the year. Amstrad's 
boss refused to commit important documents to the post because 
of the delays this entailed and badgered Mej to install a fax 
machine, then a relatively rare adornment of British offices. Mej 
prevaricated, until one day a van arrived at his office with a fax 
machine - courtesy of Amstrad. Sugar told him that the fax 
machine was a loan and he might want it back. Six years later, 
Sugar had still to reclaim his loan. 

On the software side, Chris Hall returned to his partner, 
Richard Clayton in Dorking, with the barest of specifications. 
Joyce was to be aimed at the typewriter market and have a 10- 
page manual. Locomotive Software's main job was to write the 
word processing program for the Amstrad machine. Bob Watkins 
had irritated Hall at the Brentwood meeting by emphasizing that 
Locomotive need not come up with something as good as Wordstar 
- an ageing word processing program regarded by Locomotive as 
difficult to use. 

Shortly after the meeting, Locomotive Software decided to 



170- 



ALAN sugar: the amstrad story 

clarify its client's unusually short specification. Chris Hall drew 
up a list of features which could appear in a word processor, 
classifying them under three headings: basic, useful and advanced. 
As he had anticipated, Amstrad phoned to say they wanted all of 
the basic and useful features, but none of the advanced, Amstrad 
barely interfered with Locomotive's work during the rest of the 
design phase, asserting its client's prerogatives only on minor 
details like how the end of a page of text appeared on the screen. 
The company's interventions were invariably prompted by its 
anticipation of how customers in a store would react to a particular 
feature. 

Locomotive itself prompted the only major change to Joyce's 
specification made during the design phase. The original plan was 
for the word processor to have 128k of memory, but the Locomo- 
tive team became convinced that it would be a much better 
machine if the memory could be doubled to 256k. Chris Hall 
buttonholed Alan Sugar at an event which the two were attending 
and explained the case for increasing Joyce's memory. It took 
Sugar two minutes to agree: the declining price of memory chips 
reduced the cost implications of the change. Tt was the best two 
minutes I spent on the project,' Hall says. 

In the event Locomotive's word processing program, which it 
called Locoscript, was welcomed as easy to use by most reviews in 
the computer press after the launch of the PCW8256. However, 
the software company repeated the mistake it had made with the 
CPC464 by not pressing for a royalty payment on the sale of each 
copy of its software. Instead, it agreed a fixed price of £75,000 for 
writing Locoscript. 

The price envisaged for Joyce was the aspect of Sugar's plans 
which most astonished Locomotive Software and MEJ Electronics. 
Joyce was to be a professional business computer, in the sense that 
it had enough memory to run dedicated word processing and other 
business programs, yet it would retail at under £400. The contrast 
with the word processor which the same individuals had helped to 
design for Data Recall was striking, as Locomotive's Richard 
Clayton notes: Tt was potentially the same sort of machine we'd 



171 



DAVID THOMAS 

been building four or five years before, but it was going to cost 
£399 rather than £10,000. That was absolutely stunning/ 

The low cost was achieved in two main ways. First, by integrat- 
ing the electronic controls in the way outlined by Sugar in the 
'tablets of stone'. People who opened up a Joyce were astounded 
by how 7 few chips were embedded in its unusually small printed 
circuit board. The main board contained just 17 chips, including 
eight 256k memory chips and a large gate array - the densely 
packed chip which was specially designed to carry out the precise 
functions of the word processor and which saved money by doing 
away with the need for dozens of discrete components. Crucially, 
Joyce's printer ran on the same set of controls as the central 
computer. Mej describes the design: 'We managed to get almost 
all the electronics to be inside the gate array, which makes it so 
much cheaper. We integrated all the controls of the printer into 
the main board/ 

The concentration of functions within the gate array was helped 
by the freedom which Amstrad allowed its designers. Mej elabor- 
ates: The great thing about the 8256 was that we didn't have to 
be compatible with anything. That gave us an immense amount of 
freedom. It's quite unusual to have that sort of brief from your 
client: forget about the existing standards.' 

Amstrad's technique of ordering parts in very large quantities 
was the second major way of driving down the cost. Mej was told 
to think in terms of an initial order for 100,000 Joyces, a volume 
which he regarded at the time as being too optimistic. He 
canvassed guide prices from the various parts manufacturers. Alan 
Sugar and Bob Watkins would then close the deal and Mej was 
constantly surprised at the amount the Amstrad pair shaved off 
the guide prices. 

Mej explains one of the ways in which Amstrad cut prices to the 
bone by placing large orders: 

In very large volumes, you begin to start talking prices that 
reflect the real cost of a part. Take gate arrays as an example. 
Gate arrays often reflect the latest technology, so if you buy in 
small quantities then the manufacturers try to recover the 



172- 



ALAN sugar: the amstrad story 

development costs of that technology in their prices. Amstrad 
was able to persuade many suppliers to get back their develop- 
ment costs from other people. Amstrad said it was going to give 
them a very large order and in return it wanted a cost which 
reflected the materials, not the development cost. The point was 
Amstrad could place its huge order elsewhere and the manufac- 
turer would still have to recover its development costs. 

The instruction manual proved to be the one major feature which 
was botched during the development work; after the PCW8256's 
launch, it attracted many user complaints. Alan Sugar blames 
boffin culture for the manual's faults: 

I kept hearing these rumours of people not being able to 
understand the instruction book. So one day I said, 'Give me 
the book and I'll have a look at it myself Then I realized that 
it had been written by a bunch of engineers. I looked at it and I 
thought to myself, There's no bloody way a mug like myself 
would know how to use this.' It assumed that people knew 
about high density disks and formatting disks and all that 
cobblers. We had made certain assumptions which we should 
never do. We had assumed that the engineers knew our philos- 
ophy by now. We told them to write the book. And of course 
they all sat down when the book was finished, the intellects 
together, and appreciated the excellence of the book. They all 
thought they were going to get the Nobel prize. From an 
engineering point of view, the book was fantastic. It had cross 
references for every single detail of the Z80 processor. But that 
doesn't tell somebody how to write a quick letter about selling a 
lawn-mower. 

Locomotive Software, which wrote the manual, admits to one 
basic mistake in the instruction book: 'You have to read it from 
one end to the other to understand it. Of course, everybody who'd 
been working on the project had read it from one end to the other. 
But it was impossible to dip into,' acknowledges Richard Clayton. 
But Locomotive had also operated within a formidable set of 



173 



DAVID THOMAS 

constraints which flowed from Amstrad's cost-cutting ethos. The 
PCW8256's instruction book was printed in South Korea, because 
publishing costs were cheaper there. Unfortunately, it takes longer 
to produce a book than to duplicate the software for the word 
processor, which meant that the proofs for the instruction book 
had to be approved before the master codes for the software, in order 
to have the books and the disks ready at the same time in Korea for 
shipment to Britain. As a result, parts of the book were written 
before the feature to which they referred had been finalized. Roland 
Perry, the PCW's project manager, recalls the problem: 

When the part of the book which described unpacking the 
computer had to be finished, we hadn't even designed the foam. 
We didn't know what the packaging was going to look like. 
Similarly, the software design was in such ferment right to the 
end that we didn't know which things were going to be included. 
For example, decimal tabs [a way of lining up figures automati- 
cally along a decimal point]: we didn't know whether they were 
going to be got working in the three days we had left before 
sending off the master disks. 

Amstrad learned the lesson for later products. In future, w T ork 
continued on the manuals until the boats from the Far East 
carrying the first batch of computers were a week away from 
Britain. A local printer would then rush out about 5,000 manuals, 
which would be placed into the cardboard boxes containing the 
computers at Amstrad's Shoeburyness facility. The same would 
occur with disks. Once the initial batch of products had been 
distributed, the masters of the books and the disks would be 
despatched to Korea ready for large volume production a month 
or two after the launch. This procedure involved the company in 
some extra cost, but it bought invaluable extra development time. 
Amstrad initially wanted the design work wrapped up by May 
1985, but it was not finally completed until July - a reasonable 
degree of slippage for such a project. One factor behind the delay 
was that Locomotive and Mej had been commissioned to design 
not one but two new computers. Sugar had outlined the second 



174* 



ALAN sugar: the amstrad story 

machine on page 2 of the 'tablets of stone' fax, where he described 
it as 'a full colour computor'. 

In one sense, the colour computer was to be an upgrade of the 
CPC464, because CPC software would run on it: hence its 
codename of Ant (Arnold Number Two). But it was also intended 
to be compatible with Joyce. Indeed, Ant could almost be 
described as the colour version of Joyce. The two were to be 
launched at the same time and users would decide whether they 
preferred Ant's jolly colour or the printer which only came with 
Joyce. The design work on Ant was fairly advanced when the 
project was suddenly cancelled, Amstrad having concluded that 
its designers were too stretched to complete both projects in time 
for the autumn 1985 season. It was also becoming increasingly 
clear that Joyce was the machine with the potential to turn the 
computing market upside down. 



The success of the word processor was perhaps the clearest 
example in the 1980s of the 'Amstrad effect' - the creation of a 
whole new market by supporting an innovative, keenly priced 
product with intelligent well-targeted marketing. The PCW8256 
was unveiled with what was becoming typical Amstrad fanfare in 
August 1985 in a conference centre near the City. Three actresses 
represented three different types of secretary: a frightfully snooty 
one would not dream of using anything less than a £10,000 word 
processor; a tarty secretary swore that a typewriter was good 
enough for her; while the cool, efficient secretary, of course, 
preferred the Amstrad word processor. 

Alan Sugar was on form at the launch, predicting that his new 
product 'will blow the lid off the personal computer and word- 
processing market'. He foresaw wide demand for his machine: 'We 
have brought computing and word-processing within the reach of 
every small business, one-man band, home-worker and two-finger 
typist in the country - not to mention the company chairman who 
wants one for himself, his secretary and all his managers.' 

Amstrad's chairman did not pass up the opportunity to take a 
swipe at the opposition. Claiming that his word processor was 



175 



DAVID THOMAS 

Tour times cheaper than the nearest thing we know', he added, 
'The PCW8256 costs less than the average electric typewriter and 
yet it has features that will make the big transatlantic names 
wince.' 

A few techno-purists sniffed at the machine's 'obsolete' tech- 
nology - its Z80 processor, 3" disks, television tube and so on. 
Sugar reserves a special contempt for techno-snobbery: 

The pundits thought we were bloody mad again. We were using 
the Z80 processor. When we showed it to the pundits, they all 
laughed. They looked at us as the poor relation who needed to 
go to the mental asylum. We had some very funny comments 
from the snobs in the market about the Z80: 'Doesn't Mr Sugar 
know there is such a thing as an 8086 or even a 286 processor 
available for such applications?' What they'd missed is that the 
people who bought them didn't give a shit whether there was an 
elastic band or an 8086 or a 286 driving the thing. They 
wouldn't know what you were talking about. It was bringing 
computing to people who never even thought they would use a 
computer. 

Most commentators were in fact ecstatic about the PCW8256. It 
was difficult to be otherwise with a machine which provided 256k 
of memory, a printer, disk drive and monitor, word-processing 
software, together with a package of business software - all for 
£399 plus VAT. Computer journalist Gliy Kewney said that the 
Amstrad machine would spare him much trouble at parties. In 
the past, he had been obliged to embark on a long disquisition 
when asked to recommend the best word processor, but in future 
he would be able to say 'the Amstrad' and then go and have a 
drink. 'Truly amazing for a computer in this price range,' con- 
cluded Which Computer? 'Nothing else comes close,' agreed Popular 
Computing. 'Spectacular value for money,' chipped in Personal 
Computer World. One of the most perceptive comments in the 
specialist press was made by Electrical Retailer and Trader, the trade 
magazine which described the 8256 as 'a grown-up computer that 



176 



ALAN sugar: the amstrad story 

does something people want, packaged and sold in a way that they 
can easily understand, at a price they'll accept.' 

At the word processor's launch in August 1985, Alan Sugar said 
that 40,000 machines a month would be pouring out of South 
Korea by October. These arrangements would secure an ample 
supply of the new computers. He had also been busy ensuring that 
the machines were greeted with a groundswell of demand when 
they reached British shores. 

Mass advertising through the popular newspapers and television 
had been central to Amstrad's marketing strategy ever since it had 
helped make the Tower System so successful. Spending large 
amounts on advertising generated the volumes which were essen- 
tial to Amstrad's low price strategy. But the company realized 
that the nature of its advertising would have to change as it pushed 
deeper into computers. 

Amstrad's computers were not aimed primarily at the truck 
driver and his wife or the C2, DE market generally - but at the 
small business person, the professional, the individual who worked 
from home. A different sales pitch was needed for this AB group 
of consumers from that evolved by Amstrad for its audio products. 
Malcolm Miller, Amstrad's marketing director, was instrumental 
in piloting this change in strategy through the company. 'I felt - 
and Alan didn't always agree - that with our computers, we 
needed an advert that talked to business people, not just to the 
average bloke in the street.' 

Amstrad's advertising had been notable for its lack of frills; it 
told consumers what products were on sale, how much they cost 
and where they could be bought . . . period. Sugar felt comfortable 
with this style of advertising, which chimed with his instinctive 
dislike of anything pretentious or opaque. But Miller argued that 
Amstrad's new target customers needed a subtler approach; ABs 
would be bored by the company's traditionally deadpan commer- 
cials, so these would have to become more sophisticated, even 
witty. Not only was quality advertising the way to sell computers, 
it would also help to bury the cheap and cheerful street-trader 
image still clinging to Amstrad. In short, it would do wonders for 
Amstrad's corporate profile. 



177 



DAVID THOMAS 

Eventually persuaded of the sense of this argument, Sugar 
agreed that Amstrad should find a new advertising agency to 
promote its computers, leaving audio and other products in the 
hands of its previous agency, Rupert Lovell Curtis. In late 1984, 
three agencies were invited to pitch for Amstrad's computer 
business, including Delaney Fletcher Delaney, a fledgling group 
which had recently been formed by some high flyers breaking 
away from larger agencies. 

Malcolm Miller briefed Greg Delaney before the presentation, 
advising him to cut out the detailed analysis of target markets 
which advertisers typically make when they pitch for business. 
This flannel would bore Alan Sugar, since no one knew more 
about the consumer electronics market than the Amstrad chair- 
man. Instead, Delaney should come straight to the point by 
talking Sugar through sample advertising scripts for Amstrad's 
computers. 

Delaney took Miller's advice, presenting three scripts comfort- 
ably within an hour. After he had finished, Sugar tore apart the 
first script; admitted that the second one was reasonable, but then 
explained its inadequacies; and said the third was a bit more like 
it. End of meeting. Delaney was not to know at the time, but this 
was high praise from the Amstrad chairman, who rarely descends 
to ad language to describe the efforts of those who work for him. 
Few things are brilliant, amazing or even good. They are, at best, 
O.K. This is one of Sugar's ways of keeping people on their toes. 

Delaney Fletcher Delaney won Amstrad's computer account, 
beginning their work with the CPC series. 'Who says business and 
pleasure don't mix?' was the agency's way of conveying that the 
CPC6128 straddled the games and business markets. But the first 
big splash the agency made for Amstrad was over the PCW8256. 

Taking its cue from Amstrad's attack on the typewriter market, 
Greg Delaney shot an ad with dramatic scenes of a truck dumping 
a pile of typewriters into a scrapyard full of similar dead machines. 
After describing the PCW8256, the commercial rounded off with 
the slogan: Tt's more than a word processor for less than a 
typewriter.' The advert was typical of the aggressive material 
which Delaney Fletcher Delaney was to prepare regularly for 



178 



ALAN sugar: the amstrad story 

Amstrad, but the agency had to steer it through two very different 
groups of people before it could be shown on air. 

The first hurdle was the Independent Television Companies 
Association (I.T.C.A.) which vets all television commercials before 
they are broadcast. I.T.C.A. at first rejected the advert out of 
hand, as Delaney recalls: They said that you're claiming it's 
cheaper than a typewriter, when it's not. There are typewriters 
which are cheaper. You're also showing typewriters being thrown 
away, and you're not allowed to rubbish the competition - "ash 
can advertising", as it's called in the business.' 

Delaney had to write the advert around these problems. He 
inserted the word 'most' into the slogan, which eventually read: 
'It's more than a word processor for less than most typewriters.' 
The agency ensured that all the typewriters pictured in the 
scrapyard commercial were obviously old machines. It then 
argued that the commercial was designed to persuade white-collar 
workers to move on to a word processor when they were replacing 
clapped-out typewriters, and that it did not rubbish manufacturers 
of new typewriters. This was enough to appease the I.T.C.A. 

Amstrad was the second group of people who needed convinc- 
ing. Sugar wanted to cut down on the moody scrapyard shots in 
favour of more solid information about the word processor. The 
film production company working on the commercial was aston- 
ished one day when he walked in and asked to see the rushes. 
Although it is standard practice for the chairman of a company 
selling consumer goods to take a close interest in the final shape of 
adverts, no one at the production company or at Delaney Fletcher 
Delaney had ever heard of a chairman sorting through raw film 
footage. Sugar wanted to know precisely what film had been taken, 
so that when it came to editing he would be in full control of 
decisions about which shots might be added or deleted. 

Amstrad's roll-out of a computer product tends to describe a 
standard trajectory: launch at a computer fair, followed by adver- 
tising in the trade and national press, followed three to four 
months later by a big splurge on television. Invariably spending 
large amounts on television advertising, the company was the first 
or second largest spender in its category in 1986, 1987 and 1988, 



179 



DAVID THOMAS 

according to the survey of Britain's top 100 advertisers published 
annually by Campaign^ advertising's trade magazine. Yet Amstrad 
does not spray money at a product in random fashion. It closely 
dovetails the size and timing of its television advertising both with 
consumer reaction to the initial burst of press advertising and with 
the flow of machines from its suppliers: ; It's true in a sense that 
Alan is a risk-taker, but equally he doesn't half cover himself. He 
makes sure there is a market there. He doesn't just jump in,' notes 
Greg Delaney. 

Sugar was also careful to ensure that the word processor was 
well received by his customers, the retailers who were the channel 
between Amstrad and the end users. A few months before the 
launch of the word processor, he took a prototype to show Stanley 
Kalms, chairman of Dixons, the high street electrical goods chain. 
From the start Kalms had a gut feeling that this product would be 
a winner, but he was also nervous about committing himself to 
something which had no precedent: 'O.K., it looks good for 400 
quid, but who is going to buy the damn thing?' he asked. 

Sugar took Kalms through Amstrad's thinking about pricing 
and target audience and must have been persuasive, because 
Kalms placed an initial order for 20,000 word processors. But the 
Dixons chairman - an equally canny negotiator - insisted on a 
quid pro quo for what was then a very large order for both firms. 
He asked for Dixons to be given the exclusive right to sell the word 
processor in the high street for an initial period. 'It was typical of 
the way we would buy. If we stuck our neck out for 20,000 pieces, 
we wanted a short competitive edge, a small exclusive window,' 
Kalms says. 

Sugar agreed to grant Dixons exclusivity for four months until 
the end of 1985. Amstrad would in any case have found it difficult 
to supply many more machines in addition to Dixons's order 
before the end of the year, because it would take several months 
for the supply of word processors to build up to peak volumes, 
allowing for shipping times from the Far East. But Sugar also 
knew that the arrangement with Dixons would bring Amstrad 
many advantages. Having one main outlet for a new product 
simplified the initial sales message: Amstrad need do no more than 



180 



ALAN sugar: the amstrad story 

tell people to go to a Dixons store. If sales through Dixons went 
well, other stores would be stung into taking large volumes of the 
word processor once the Dixons exclusivity ended. Moreover, the 
arrangement would encourage Dixons to make a big push for the 
product, as Sugar explains: 'When Dixons have a gut feeling for a 
product, they get behind it themselves. They will stack it up, pile 
it high, display it in all the stores, give it prominence, spend a bit 
of their own advertising on it.' 

The reaction of Dixons's customers to the word processor 
surpassed the highest hopes of Sugar and Kalms, as the Dixons 
chairman recalls: 

It was one of the great phenomenal take-offs in my experience. 
Products very rarely take off at the beginning. Usually, they're 
on the market years before they take off They're like actors: 
they can be around for years and one day they become stars. 
But this was an absolute bombshell. He'd brought a word 
processor and a computer down to the level of your school- 
teacher, your local vicar - everybody who was in the business of 
typing and communicating. Till then, home computers had been 
just sophisticated toys. Here was the first real professional piece 
of equipment. So every charity in the country wanted one. Every 
secretary in every small company went to her boss and said: T 
want one.' 

Before the appearance of Amstrad's word processor and its 
subsequent line of I. B.M. -compatible personal computers, pro- 
fessional computers had been sold almost entirely through special- 
ist dealers. Amstrad and Dixons were the main catalysts in moving 
sales of professional computers into Britain's high streets. Since 
the early 1980s Dixons had been trying to add personal computers 
to its traditional lines such as cameras and hi-fi; it had sold a 
range of computers made by Japanese companies like Epson, 
Sanyo and Toshiba, but without much success, as Kalms notes: 
'We had been in computers three or four years, making no money, 
persisting as you often do with a new product group, but waiting 
for the point of entry. We saw that computers had to be a mass 



181 • 



DAVID THOMAS 

consumer product eventually. And Alan made them a mass 
consumer product.' 

Dixons had been looking for a computer which could be sold 
without needing specialist advice about peripherals and software. 
Amstrad filled that gap. The dream of a business like ours was 
that computers should cease to be exclusive, magic products. They 
would become consumer electronics. That was the great break- 
through. That was the measure of Alan's genius,' Kalms says. 

Amstrad's pricing policy was fundamental to this breakthrough, 
for Sugar believed that the price of his computers should reflect what 
they cost to make, not what the market would bear. He differentiates 
this approach from that of the established computer companies: 

The big computer companies were ripping people off for a long, 
long time. £2,000 was the price of a personal computer and that 
was it. It bore no relations to the bill of materials, but that's 
how they made their money. We came along and said it's about 
time we broadened the market. We looked at a personal 
computer and we said this is a £399 piece here - though it was 
a bit tight at £399. 

Some retailers tried to persuade Sugar to increase the price: 
Teople said to us, "You're bloody mad. Absolutely mad. Why 
don't you sell it for £499 or £599? It doesn't have to be £399.'" It 
was a plea which he resisted. 

Thanks to the elegant simplicity of the Amstrad word processor, 
its unbeatable price and the thoroughness of the company's 
marketing plans, 1985-86 proved to be an annus mirabilis in 
Amstrad's history. Some inkling of what was to come emerged in 
February 1986 when Amstrad released figures for the first half of 
its financial year, the six months to December 1985. The first two 
sentences of Lex's comment in the Financial Times caught the tone: 

The City may have some difficulty understanding anything to 
do with electronics, but rarely has it misjudged a company so 
resoundingly as Amstrad. While the stockbroking fraternity was 
losing its collective sleep on audacious forecasts of a 50 per cent 



182 



ALAN sugar: the amstrad story 

increase in pre-tax performance, Amstrad quietly announced 
just short of a trebling of interim profits - from £9. 5m to £27. 5m. 

Amstrad's share price had begun to climb steeply in the autumn 
of 1985, after the well-received launch of the PCW8256 and the 
release of financial results demonstrating the success of the 
CPC464. The stock price jumped a further 50p to 342p on the day 
in February 1986 when the half-year results were announced. In 
twenty-four hours the value of Sugar's paper fortune increased by 
£27.5 million to £188 million. 

Sugar took the opportunity to add to his real fortune and in May 
1986 raised £25.7 million by selling another 5 per cent of Amstrad 
stock, further reducing his stake in the company to 45.4 per cent. 
Given Amstrad's performance, there was no problem in placing the 
5,000,000 shares at 520p each with institutional investors, some with 
existing stakes in Amstrad and others new to the company. 

Amstrad's chairman issued a formal statement explaining the 
share sale: 'It seemed to me to be about time that I reduced my 
holding in Amstrad in order to develop my interests outside the 
company. But Amstrad will still absorb all my management 
attention.' The Amstrad chairman gave a more characteristic 
answer when asked by Microscope computer magazine what the 
money was for: Tt's money for Al. I'm not starting a new company. 
I'm not building electric cars [the Sinclair electric car fiasco was a 
recent memory], I'm not funding any new developments. It's my 
own money.' 

Amstrad's surge forward was confirmed in the full year results 
for 1985-86, issued in October 1986. Profits increased by an 
astonishing 273 per cent to £75.3 million on sales which were up 
123 per cent at £304.1 million. The shares - which had been split 
in five again in June - reached 136p, valuing Sugar's stake in the 
company at £340 million. The stock market now reckoned 
Amstrad to be worth more than Ferranti, a long-standing pillar of 
the British electronics industry. One brokers' analyst commented 
that, considering how few assets the company possessed, Sugar 
was probably the most highly capitalized individual on the London 



183 



DAVID THOMAS 

stock market. In both 1985 and 1986, Amstrad was among the top 
three performing shares on the market. 

Amstrad had not neglected its traditional products during its 
launch into computers. The company re-entered the video recorder 
market with a machine imported from Japan in 1985, and by June 
1986 was claiming 10 per cent of the British market. In January 
1986 they broke new ground by embedding a compact disc player 
in a hi-fi system which cost less than £300. 'We see CD as the 
future in sound reproduction, but, until now, [it] has been only 
affordable by the rich,' Sugar said when announcing the move. 
Philips, the Netherlands-based electronics giant, was stung into 
revamping its compact disc player strategy a month later. ' If you 
have Amstrad shares, sell 'em,' quipped Leigh Robinson, Philips's 
marketing manager in the U.K. 

That almost everyone ignored this advice was undoubtedly due 
to Amstrad's computer activities rather than to anything it was 
doing in audio or video. Its impact on the word processing market 
turned out to be as revolutionary as Sugar had predicted. Before 
the launch of the Amstrad PCW8256, word processor sales in 
Britain were bumping along at 50,000-75,000 a year. But Amstrad 
sold 350,000 word processors during its first eight months in the 
market, attracting a whole new generation of users to computers 
for the first time. 

Amstrad was the most visible beneficiary of this word processing 
revolution, but its entry into computers caused ripples throughout 
the microcomputer industry. Take software as an example. 
Amstrad's word processor, together with the low cost I.B.M.- 
compatible machine it launched in September 1986, helped to 
transform the British business software industry. It had been used 
to supplying high cost, low volume products and a business 
software package would typically have cost more than the retail 
price of an Amstrad computer. After Amstrad's entry that no 
longer made sense, as software entrepreneurs soon realized. 

David Goldman, founder and chairman of Sagesoft, first saw 
Amstrad's PCW8256 when it was launched at a computer show in 
August 1985. Based in the north-east of England, Sagesoft had 
been prospering in a modest way since it was founded in 1981. 



184 



ALAN sugar: the amstrad story 

Goldman immediately recognized the opportunity offered by 
Amstrad's machine, as he explains: 'It just seemed pretty obvious 
to me that it was going to make a huge impact on the market.' 

Goldman's reaction was to rush back to Newcastle and modify 
an accounts package which Sagesoft had devised for CP/M and 
which had been selling for £375. Within a month Sagesoft had a 
slimmed-down version on sale, suitable for Amstrad's word pro- 
cessor and with a price tag of £100 including VAT. Before the 
PCW8256 launch, Sagesoft had been selling about 250-300 units 
of this software a month. By the start of 1986, monthly sales had 
shot up to 5,000. 

Sagesoft was to repeat the trick after the well-trailed launch of 
Amstrad's I. B.M. -compatible personal computer. Goldman adver- 
tised spreadsheet, word processing and other software specially 
modified for this machine a matter of days after Amstrad's own 
advertising campaign began. For a couple of years after the launch 
of the PCW8256, Goldman says, Amstrad was the most important 
factor spearheading the development of mass market software and 
therefore the prosperity of companies like his own. His company, 
by then known as Sage Group, was valued at £21.1 million when 
it was launched on the stock market in December 1989. 

Amstrad prompted a similar upheaval in specialist office equip- 
ment retailing. Some specialist computer outlets suffered because 
they were slow to realize that it was ushering in an era of low-cost, 
mass market machines, but those who understood the changes 
prospered. While Amstrad's main sales thrust for its computers 
was through high street chains, it did not neglect the more 
specialist outlets, which were exempted from the terms of its 
exclusivity arrangement with Dixons. Terry Wilding, an East- 
Ender like Sugar, had launched Wilding Office Equipment in the 
mid-1960s, and by 1985 it was heavily dependent on the traditional 
typewriter market. Until then, Wilding's dealings with Sugar had 
been as a supplier of office equipment to Amstrad. 

Before the word processor's launch, Sugar told Terry Wilding 
that he had a product which might interest him. Wilding recalls 
that he was impressed with the prototype of the PC W8256, but - 
seeing no reason to believe it would take off - he took only a few 



185 



DAVID THOMAS 



for his shop windows. Customer interest was so intense, however, 
that he put in a large order. Soon he was selling more than 1,000 
PCW8256s a month, significant volumes for what was then a 
relatively small group. 

During the next couple of years, Amstrad played the same 
catalytic role in Wilding's growth as in Sagesoft's. 'Amstrad helped 
to take us out of being a traditional office equipment supplier into 
computers. We did very well out of them,' says Wilding, whose 
group was valued at over £12 million when floated on the Stock 
Exchange in November 1986. 

Some of the factors underlying Amstrad's performance in 
1985-86 were as remarkable as the headline figures such as the 
near tripling of profits. Amstrad's profit margins increased from 
14 to 25 per cent in the year to June 1986, an astounding figure for 
a manufacturing company battling it out in the cut-throat world 
of high street consumer electronics. Luck played a part in the 
achievement of these profit margins, since its launch into com- 
puters coincided with a cyclical downturn in the price of memory 
chips, a notoriously volatile commodity whose supply swings from 
feast to famine and back again in short order. But Sugar's 
management of Amstrad's growth also underpinned his ability to 
squeeze such high profits out of the company. In the financial year 
1985-86, when Amstrad's sales grew by 123 per cent, the company 
increased its work-force by a mere 6 per cent, adding only thirty 
people to its existing complement of 496. On average, each 
Amstrad employee generated sales of more than £500,000 in the 
year - an astonishing ratio. 

A tight ship, spiralling sales and little demand for capital 
investment - it was a recipe for generating cash. As Amstrad grew, 
it had to finance ever larger orders from its suppliers, which forced 
the company into substantial borrowings of a temporary nature 
while waiting for revenues from its sales to flow in. Yet while the 
scale was bigger, the underlying pattern of its finances had not 
changed greatly from the 1970s. Over the course of a year, 
Amstrad was able to fund its activities out of its cash flow and pile 
up cash in the bank. By the end of 1985-86, it was sitting on a 
cash pile of about £50 million. 



[86 



ALAN sugar: the amstrad story 

Confident that his entry into computers had been well judged, 
Sugar felt justified in articulating Amstrad's philosophy for the 
computer market in his chairman's statements in 1985 and 1986. 

Point 1 was on design: 'We are experts in design and engineering 
- one of our talents is to engineer products with all the specifica- 
tions and facilities the market demands and delete those unused 
facilities that are only enjoyed by the minority. In short, we 
produce what the mass market customer wants and not a boffin's 
ego trip.' 

Point 2 was on manufacturing flexibility: 'Now that the sales 
volumes are getting larger and the market horizons wider, we are 
also broadening manufacturing sources. We procure our products 
and components from [the] U.K., Japan, U.S.A., Korea, Italy, 
Germany, France, Belgium, Taiwan and Hong Kong. We have 
the facility to expand our manufacturing capability and at the 
same time remain flexible, both from a currency exchange and a 
political point of view.' 

Point 3 was on marketing: 'Our marketing skills and under- 
standing of the seasonable nature of our business is one of the 
main reasons for our success. Having the right inventory at the 
right time is important, and making sure that our commitment for 
inventory is continually in line with our regularly updated sales 
forecasts.' 

Sugar rounded off his 1986 statement by telling Amstrad's 
shareholders to expect more fireworks from the company: 'The 
year I have just reported has shown excellent growth of sales and 
profits, once again I would like to say that this is no fluke or flash 
in the pan. The electronics industry is a fascinating one, inasmuch 
as there are always new products to develop and markets to 
explore. . . .' 

Shortly before penning the 1986 statement, Sugar gave an 
interview to International Management magazine in which he articu- 
lated his philosophy of total product flexibility. 'We're interested 
in the mass-merchandising of anything,' he told the attendant 
journalist. Tf there was a market in mass-produced nuclear 
weapons, we'd market them too.' 



• 187 



• 10- 

The Boffin and the 
Entrepreneur 



Alan Sugar knew that something big was about to break when he 
took an unexpected call from Mark Souhami, managing director 
of Dixons, early in 1986. He was in Hong Kong on one of his 
regular trips to the Far East, and the last thing he was anticipating 
was to be contacted by the powers that be at Dixons. 

'We heard you were in Hong Kong. There's something we'd 
like to talk to you about. Why don't you pop over and see us? 
We're staying at the Mandarin,' Souhami said, explaining that 
Stanley Kalms, Dixons's chairman, was also there. 

Intrigued, Sugar paid the seven-pence fare for the ride across 
the stretch of water separating Kowloon from Hong Kong Island 
on the Star Ferry. Idly watching as the ferry weaved in and out of 
the barges from Communist China heading for the open sea, he 
turned over in his mind what Dixons might want. 

However, the Dixons executives were in no hurry to enlighten 
Sugar once he had found his way to Souhami's suite in the 
Mandarin Hotel. They spent a good ten minutes in small talk 
discussing routine matters which could easily have waited until 
they were all back in Britain. Sugar's patience, never his strong 
suit, snapped. 'Come on, what the bloody hell are you driving me 
mad for? I've come over here on the Star Ferry, sweating hot, 
because you asked me. Now what do you want me here for?' 

The Dixons executives got down to business. They wanted to 
discuss a deal which could be in the offing with Sir Clive Sinclair. 



188 



ALAN sugar: the amstrad story 

After leaving the meeting with Dixons, Sugar pondered what he had 
been told. As he understood it, there was a large stock of Sinclair 
computers up for sale. Dixons had considered buying them, but the 
trouble was that Sinclair wanted to sell them as part of a package 
deal by which someone would take his entire computer business off 
his hands. Sugar was now left with the question: would Amstrad be 
interested in buying the Sinclair operation? 

Sugar already knew that Sinclair Research, Clive Sinclair's 
home computer company, had been able to limp through the 
second half of 1985 only because Dixons had bought 160,000 of its 
computers and pocket televisions at rock-bottom prices. That £10 
million deal had been enough to earn Sinclair a respite from his 
bankers, Barclays and Citibank, and creditors, led by his main 
suppliers Timex, Thorn E.M.I, and AB Electronic. However, it 
was evident to Sugar that this group wanted to sort out the 
financial problems at Sinclair Research once and for all now that 
the Christmas season was over. 

Sugar explains how he saw the line-up: 

Dixons had just finished raping Clive in the Christmas of 85. 
They were really licking their lips because they had seen all 
Sinclair's computers flowing through their shops. Clive might 
be losing his trousers on them, but Dixons were making lots of 
money from them. Then Dixons heard that the banks and the 
creditors wanted a solution by the end of March. And there was 
Dixons saying to itself that this little gravy train was going to 
end. No way could it allow Sinclair to go down the pan. 

The Amstrad boss now believed that Dixons themselves had 
thought about taking Sinclair over. Looking back at the event, 
Sugar describes his view of Dixons's position: 'Dixons began to 
realize in no uncertain terms that this was a very, very, very big 
can of worms, that involved lots of stock lying around all over the 
place. Timex, Thorn E.M.I, and AB Electronic - every bloody 
company in the world was in trouble with Sinclair. So Dixons 
decided to drop it on me instead.' 

Sugar was more than happy to take up the running, because he 



189 



DAVID THOMAS 



could smell the potential. Despite its recent troubles, Sinclair was 
still the dominant force in games computers and accounted for 
some 40 per cent of the total U.K. home computer market. By 
acquiring the business, at a stroke Amstrad would become the 
market leader in games computers, boosting its total computer 
sales by hundreds of thousands of units a year. It would add 
another string to Amstrad's bow at the entertainment end of the 
market, while Sugar pursued his main strategy of pushing further 
into business computing. Moreover, Sinclair had sold over 5 
million computers including more than 1 million Spectrums, its 
flagship product. Sugar felt he could exploit this huge installed 
base by selling computer peripherals to Sinclair users. 

The Amstrad chairman was confident that he could squeeze a 
lot more money out of the operation than Clive Sinclair. He would 
improve the efficiency with which the Sinclair machines were 
made, eliminate overlapping distribution costs and inject Amstrad 
zest into selling them. From his point of view, the deal would be 
all but perfect if he could also win the exclusive right to the 
Sinclair name, one of the most powerful brands in Britain. 

Already excited by the deal's potential, Sugar called a meeting 
for 9 o'clock the next morning at the Amstrad offices in Kowloon. 
Bob Watkins was there, as were the people from Dixons. Sugar 
took charge of the proceedings, wheeling in one of the company's 
young Chinese designers. Following instructions from Sugar and 
Watkins, the designer sketched out an improved version of the 
Spectrum Plus, the latest Sinclair machine, featuring a keyboard 
with a cassette mechanism tagged on to it. 

'Is this the kind of thing you'd want to buy?' Sugar asked 
Souhami. 

'What kind of price could it sell for?' countered the man from 
Dixons. 'About £159/ Sugar replied. Dixons were more than 
interested. 

Sugar went into overdrive, phoned Amstrad headquarters in 
Brentwood and told them to have a Sinclair Spectrum delivered 
from the local Dixons shop immediately. An Amstrad engineer in 
Brentwood opened up the machine and shouted down the phone 
to Bob Watkins a list of all the components he found embedded in 



190 



ALAN sugar: the amstrad story 

the printed circuit board; he then faxed through to Hong Kong 
the Spectrum's circuit diagram. By the next day, Sugar and 
Watkins had a provisional estimate of how much it would cost 
Amstrad to make a Spectrum, based on that list of components 
shouted over the phone. 

Sugar explains his thinking: 'We knew how much a cassette 
mechanism cost and we took a flying guess at a few other parts 
that were in the Sinclair. And we concluded that if we pressed the 
button, we could tool up for that product, make 400,000 and have 
them in the U.K. market and all the other markets of the world in 
time for Christmas 1986/ 

The Amstrad boss knew he could make a tidy profit from the 
Sinclair business: that much was clear from the favourable Dixons 
reaction to the price he had quoted for the enhanced Spectrum 
Plus, and from his calculations as to the cost of making the 
machine. He was now hooked on the project. Once back in Britain, 
he called up a supplier in Taiwan to discuss making the Sinclair 
machines for Amstrad. Amstrad's engineers in Brentwood drafted 
preliminary sketches of the manufacturing tooling that would be 
needed to make the Spectrum, and faxed these to the Taiwanese 
manufacturer. Sugar phoned Taiwan again to confirm his interest: 
There could be a deal cooking, but keep quiet. I can't tell you no 
names yet. I'll come back to you.' 

Sugar had travelled this far without even talking to Sinclair or 
Sinclair's creditors. With the end-of-March deadline fast 
approaching, however, a meeting had to be arranged. Both sides 
agreed that Alan Sugar and Sir Clive Sinclair would kick off the 
talks by meeting on their own without advisers. As a symbolic 
half-way house, they chose for their encounter a restaurant in 
Liverpool Street station, the London railway terminal which 
served both Sinclair's Cambridge headquarters and Sugar's Brent- 
wood base. 

Anyone who saw the two men lunching together that day would 
have been struck by the contrast. On one side of the table sat the 
bespectacled Sinclair, whose domed head and quiet, precise 
manner were perfectly suited to his reputation as Britain's leading 
boffin. He had projected himself as the ultimate technological 

• 191 • 



DAVID THOMAS 

guru, a self-image endlessly amplified by Britain's mass media to 
the point where the Sun described him as 'the most prodigious 
inventor since Leonardo'. At the height of his computer success, 
his Cambridge headquarters was a monument to high-tech high 
taste. An atrium featuring a bronze sculpture led off the stainless- 
steel-clad offices, while one visitor breathlessly recorded, 'Even the 
bicycle shed was fine art.' 

On the other side of the table sat Alan Sugar, a bear of a man who 
would never knowingly let fine art darken the threshold of Amstrad's 
offices in Brentwood, a town as nondescript as Cambridge is 
distinguished. The last title he would ever aspire to was that of 
boffin, which for him is a term of abuse. Unlike Sinclair, he had no 
emotional commitment to his products or the technologies that went 
into them; his unbending commitment was to the bottom line. 

Yet both Sinclair and Sugar were seen - at different times and 
for different reasons - as the great white hopes for a British 
business revival in the 1980s. The burden of carrying these hopes 
was on the point of passing from Sinclair to Sugar as the two 
Thatcherite role models ate their lunch. 

Both these electronics tycoons have warm memories of the 
occasion. Even Sinclair, who might be expected to look back at 
the events with some regret, says, 'I found Alan Sugar a delightful 
man to deal with. He tended to say - this is the deal. He never 
tried to improve his position or deviate from what he said he 
would do. He was very straightforward and clear-headed. He was 
very pleasant company, enjoyable to meet - a witty man.' 

A misunderstanding had to be cleared out of the way first, 
however. Sugar was under the impression that he was being asked 
to buy the whole of Sinclair Research, which would have meant 
acquiring all Sinclair's various research projects and could have 
meant taking Clive Sinclair into the Amstrad operation. Luckily, 
Sinclair put his cards on the table: 'Look, Alan, I'm not going to 
work for anybody.' 

Sugar heaved a sigh of relief. He viewed Sinclair as someone 
whose reputation had been over-hyped by the media. While 
admiring Sinclair for having traded on the image the newspapers 
had woven round him of genius, entrepreneur and scientist 



192 



ALAN sugar: the amstrad story 

extraordinaire, admiring a man for having played his hand to good 
effect was different from wanting to work with him. 'Look, Clive, 
you can carry on talking in your cultured Etonian accent and 
fooling everybody into treating you as an undergrad in your lab in 
Cambridge, but I have your card marked. So let's talk business. 
What's the bottom line?' Sugar asked. 

Sinclair wanted Sugar to buy his computer operations, including 
the existing stock and the rights to the continuing business, but 
not Sinclair Research, which would continue as a vehicle for 
Sinclair's research interests. As the meeting wore on, Sinclair 
mentioned the figure he had in mind as the purchase price for his 
computer operations, which was a sum way in excess of anything 
that Sugar intended to pay. 

The deal was heading into a brick wall, as Sugar recalls: 'The 
bottom line was he sees a way where there could be mutual benefit 
. . . bla-de-bla. Anyway, it was a load of bullshit. He wanted a lot 
of money. I'm not interested. Goodbye. I literally felt that at the 
end of that meeting we'd hear no more about it.' 

But Sinclair's suppliers and banks were still determined to 
arrange a deal by Easter which fell at the end of March. The 
company's position was clearly slipping in the market. With the 
Christmas sales period behind it, there was the long and unattrac- 
tive prospect of the dead spring and summer sales months to limp 
through. The timing was also influenced by a capital gains tax 
liability which Sinclair Research had. It was so large that it might 
have tipped the company over into insolvency if ever it were paid 
in full. If that happened, the creditors and the banks would have 
had to realize their assets after Sinclair had gone into liquidation. 
Obviously, in that case they would receive less money than if they 
could strike a deal while Sinclair was still a going concern. If a 
deal could be concluded by the end of the tax year, 31 March, 
then Sinclair could offset losses arising out of the deal against the 
capital gains tax liability. 

A couple of days after his lunch with Sinclair, Sugar took a call 
from an accountant representing Sinclair, asking if he could come 
to Brentwood for a meeting. The accountant's opening gambit was 
to the point: 'Look, we want a deal, what are you prepared to do?' 



193 



DAVID THOMAS 

During the course of this meeting the full scale of the suppliers' 
inventory holdings emerged. Millions of pounds' worth of Sinclair 
stock was dotted about the country, both as finished products and 
as components in various stages of assembly. A hurried round of 
meetings involving advisers from both sides was arranged. The 
plan was to hammer out the final deal in a set-piece engagement 
between Sinclair (together with advisers and creditors) on one 
side, and Sugar on the other, before the end of March. 

The talks with Sinclair's advisers had already forced Sugar to 
postpone his plans to fly to his holiday home in Florida for an 
Easter break, and he had no intention of putting off his departure 
a second time. His ticket for the 7 o'clock Concorde flight that 
evening was safely tucked into his briefcase as he walked into the 
headquarters of Barclays Bank in the City to stitch up the Sinclair 
deal. He was feeling lucky: it was Monday, 24 March 1986, his 
thirty-ninth birthday. 

He recalls the scene vividly: There was this bloody great big 
round table with all these bankers sitting there, grovelling and 
hovering around. And the bottom line was the banks were in for 
God knows what and the suppliers had piles and piles and piles of 
inventory.' 

The Amstrad boss opened up by saying that he would throw £5 
million in cash into the pot. He would then go off and talk 
separately to each individual supplier and take all their stock from 
them. He reckoned there was about £11 million worth of inventory 
in total. He might be able to strike up some relationship with the 
manufacturers present, allowing them to continue to make Sinclair 
machines - but he rather doubted it, because down the years his 
experience had been that he could not buy anything decent in the 
U.K. 'British manufacturers are crap because their quality is so 
awful,' Sugar told the gathering. 

Turning to the banker from Barclays who was leading for the 
other side, Sugar summed up: That's the deal I'm offering, but I 
haven't much time. So if you're interested, fine. If not, it's nice to 
have met you and I'll say goodbye.' 

The bankers and manufacturers across the table were less than 



194- 



ALAN sugar: the amstrad story 

impressed by this performance. 'They all started saying at once 
that this was no good,' Sugar recalls. 

Ted Merrette, managing director of AB Electronic, tried to 
argue with Sugar about quality standards. He told Sugar that 
British manufacturers like themselves - an electronic component 
manufacturer based in South Wales - could meet Amstrad's 
demanding standards because by then they had caught up with 
the Japanese. The argument was not simply of academic interest 
to Merrette or the other manufacturers present, because they 
hoped to continue producing machines for Amstrad: 'Sugar was 
buying the Sinclair name for something. Presumably he was going 
to carry on selling the products. My feeling was we could produce 
them and produce them effectively/ Merrette recalls. 

Sugar was in no mood to debate with the manufacturers or to 
listen to the bankers' protests. 'Look, that's the deal. I'm leaving 
now. See you,' he told the gathering. 

The Sinclair side asked him to wait because they might have an 
alternative proposition. 'Well, I haven't got much time left because 
I've got to catch a flight in a couple of hours,' he replied. 

The other side then asked him to leave while they considered 
their options. 'We want you to increase your cash offer to £10 
million,' he was told on re-entering the room. Clive Sinclair, who 
had taken little part in the meeting until then, started talking 
about the wonders of the technology in his Spectrum machine, but 
Sugar interrupted him. 'For God's sake, Clive, I don't care if they 
have rubber bands in them, as long as they work.' 

The Amstrad chairman considered for a further micro-second 
before turning down flat the request for extra money. In the end 
the Sinclair side accepted; they had little choice. Sitting back in 
his seat in Concorde, Sugar mulled over a good day's work; the 
deal appeared to be sewn up. But this assessment proved to be a 
little premature. 

There followed a week of intense activity as the two sides tried 
to agree on the small print before the end-of-March deadline. The 
contracts were immensely complicated because so many parties 
were involved on the Sinclair side. Amstrad's directors - Ken 
Ashcroft, Bob Watkins, Malcolm Miller and Jim Rice - moved 



195 



DAVID THOMAS 

into an office at the City lawyers who were advising the Sinclair 
team; there they placed a telephone with a loudspeaker unit in the 
middle of their temporary headquarters, so that they could discuss 
progress with Sugar in Florida. The Sinclair team occupied 
another office in the same building. The two sides either negotiated 
or discussed tactics amongst themselves from 10 am each morning 
until the early hours of the following morning, every day from 
Tuesday 25 March and over the Easter weekend to Easter 
Monday, 31 March. 

The fax machine in Sugar's holiday home in Florida worked 
overtime as he kept in touch with negotiations: They were faxing 
through to me pages and pages and pages of messages. But I was 
lucky to be in Florida. They were the poor sods having to do the 
deal,' he says. 

Two issues dominated the negotiations. First, Amstrad had to 
have watertight assurances that it was buying all the technical 
drawings and other pieces of intellectual property which would 
give it exclusive ownership of Sinclair's computers. On legal 
advice, the Amstrad team opened up by demanding to see every 
notebook ever written within Sinclair Research on the Spectrum 
project. The Sinclair side countered that this would be an imposs- 
ible request for any company to fulfil, but was doubly unreason- 
able to make of one which everyone knew to have had management 
problems. Forty-eight hours passed in haggling over the assur- 
ances on intellectual property ownership which Amstrad really 
needed. 

In the end, everyone agreed that Amstrad's requirements 
centred on control of the chip design for the Spectrum: 'The heart 
of the computer was the intellectual property rights held within 
the gate array custom chip design and the code in the ROM, the 
operating system. Without that signed, sealed and delivered to us, 
we'd have bought nothing,' Sugar explains. 

A couple of problems on the ownership of the chip design had 
to be resolved as the end-of-March deadline approached. Ferran- 
ti's name appeared on one of the design documents; some time 
and much nervous energy was expended before Ferranti agreed to 
waive any claims to ownership. More trickily, a man called John 



196 



ALAN sugar: the amstrad story 

Grant - who ran a small software house, Nine Tiles Information - 
had helped to design the Sinclair operating system. Grant was 
found to have left for an Easter boating holiday on the Grand 
Union Canal, so a motorcycle courier was hurriedly commissioned 
to ride up and down the towpath looking for his boat. The dispatch 
rider kept missing Grant, who was not located until he returned 
home on Easter Monday. Grant cannily haggled for a day over 
the price for his waiver, finally signing only on payment of about 
£25,000 by the Sinclair team. 

The second central issue in the negotiations was the mind- 
numbing task of agreeing the terms governing the payment to each 
supplier for its pile of inventory. The point of contention was not so 
much the price Amstrad was offering, but the conditions attached to 
it: for example, the quality specifications to be laid down for the 
inventory which the manufacturers would supply to Amstrad and 
the help Amstrad would give the manufacturers in assembling 
component parts into finished stock. Depending on how tightly these 
conditions were drawn, the price Amstrad was offering could be 
made to appear either fairly generous or impossibly mean. 

Negotiations intensified as the deadline approached. The two 
sides met at 10 am on Easter Monday, 31 March, worked right 
through Monday night and then right through Tuesday. Tempers 
frayed and each side formed an increasingly low opinion of the 
other. The Sinclair team concluded that Amstrad was playing a 
delaying game in order to force last-minute concessions. At one 
point, someone on Sinclair's side thought he spotted an attempt 
by Amstrad to back out of a point already agreed. When the man 
read out to Sugar his notes of a previous conversation, he was 
subjected to a barrage of ripe abuse down the telephone wires for 
his pains. 

For his part, Sugar's view of the other side also plummeted as the 
faxes hummed across the Atlantic: 'I learned then what a bunch of 
gangsters the banks are. They really are gangsters. They thought 
they were dealing with a boy who had just got bar-mitzvahed, when 
they were dealing with me. They reckoned I was someone with too 
much money and didn't know what I was doing.' 

The deal nearly collapsed many times as the exchanges on the 



197 



DAVID THOMAS 

phone between Florida and London grew increasingly angry. On 
more than one occasion Sugar told his team of directors, 'Pack 
your briefcases up. Leave the room and clear off to your wives. 
We're not buying the thing.' 

In the end, terms were agreed at 3 o'clock in the morning of 
Wednesday, 2 April, technically a couple of days beyond the 
deadline. 'We were all zombies at the end of it. It's the worst deal 
I've ever done,' one of Sinclair's bankers recalls. 

Amstrad paid Sinclair £5 million for all the rights to his 
computer products, including the perpetual right to use the 
Sinclair brand name. The company paid a further £11 million to 
Sinclair and his sub-contractors for their stock. Self-interest was 
the glue which held the deal together: the Sinclair side had too 
much to lose by letting the talks collapse, while Amstrad had 
much to gain by ensuring the negotiations were successful. The 
agreement was kept under wraps until Sugar flew home from his 
holiday. 

He landed at Heathrow on Saturday 5 April. Nick Hewer and 
Andrew Mackay, Amstrad's public relations advisers, spent the 
weekend worrying over whether the news would leak out. Hewer 
waited until the Saturday before issuing invitations to a press 
conference on the following Monday; these baldly stated that Alan 
Sugar would be making an announcement with major implications 
for the European computer market. Hewer was careful to leave 
Sinclair's name off the invitation, knowing that to do otherwise 
would have given the game away to the Sunday papers. 

The first inkling that the assembled newspaper and television 
journalists had of the deal was when Alan Sugar and Clive Sinclair 
walked together into the press conference in the Howard Hotel. 
Sugar allowed himself a quiet moment of self-congratulation as he 
sat watching the television crew powdering Sinclair's head to dim 
its reflection in the lights, but he resisted the temptation to gloat: 
'The journalists tried to wind me up in the press conference to slag 
the poor sod off. Why was he so unsuccessful and so on? I wouldn't 
fall for it. There was no way I was going to run the man down as 
he was sitting there very dignified. I thought there's no mileage in 
letting them wind me up.' 



198 



ALAN sugar: the amstrad story 

Press interest in the news was intense. It was, commented the 
International Herald Tribune, 'the most widely followed $7 million 
corporate transaction in British history'. Most comment was 
favourable. 'Amstrad's purchase of Sir Clive Sinclair's home 
computer brands is so logical it could have been conceived by an 
electronic brain,' the Lex column concluded in the Financial Times ^ 
adding that Amstrad would be able to recover the purchase price 
after just one year of selling the Sinclair machines. The Daily 
Telegraph agreed: 'The man who knew what the technology could 
do has lost to the man who knew what the customer wanted.' 

Amstrad could now claim 60 per cent of the British home 
computer market - 40 per cent from Sinclair, and the rest from its 
own share. It had also overtaken Commodore as the biggest 
volume seller of low cost computers in Europe. The Times hailed 
the deal as 'the end of an era'. The mantle of Britain's foremost 
electronics entrepreneur had passed from Sir Clive Sinclair to 
Alan Sugar. 



The same month that the Sinclair purchase was announced, a 
Cambridge-educated engineer in his late twenties called Richard 
Altwasser journeyed to Brentwood for his first day as an Amstrad 
employee. He looked curiously out of place in the Amstrad 
headquarters; thin, carefully spoken and with an intellectual 
demeanour, only his mop of fair hair stopped him from passing as 
a Clive Sinclair lookalike. 

Altwasser had in fact led the small team which designed the 
Sinclair Spectrum and Sugar was about to benefit from an ideal 
coincidence: the Spectrum's designer started working for Amstrad 
at the very moment when the company acquired all the rights to 
the machine. Yet coincidence - luck, really - was indeed what was 
involved, for Amstrad had not recruited Altwasser specifically 
because of his experience with Sinclair Research. 

Alan Sugar and Bob Watkins had decided back in 1985 that 
Amstrad had become so big it needed a well staffed research and 
development unit. They wanted top-flight engineers to help design 
Amstrad's new products, especially since they knew that the 



199 



DAVID THOMAS 

company's future lay in yet heavier involvement with computers. 
Watkins began to look for someone to lead the team, turning 
eventually to head-hunters to complete his search. 

The head-hunters recommended Richard Altwasser, who by 
then was working in the Brussels office of P.A. Technology, a 
leading group of technology consultants. Altwasser duly made the 
trek to Brentwood one Saturday in January 1986 for what turned 
into a rather unusual interview with Bob Watkins. 

Watkins kicked off by telling Altwasser that he was not even 
going to try to probe his technical ability: 'It would be a waste of 
time because I'm not qualified to do so. What we need is someone 
with a brain. We intend to develop a huge number of new 
products, so we want someone who is really good and really 
understands computers.' 

He went on to explain that Amstrad was worried that it was 
becoming too big to rely totally on the small contractors like MEJ 
Electronics which had designed its computer products until then. 
'They might go down the pan one day and then we're stuck.' 

Altwasser was being offered the chance to set up an in-house 
development team by Amstrad, a company whose track record in 
taking the home computer market by storm he greatly admired. 
The Cambridge engineer explained to Watkins the number of 
people, the floor space and the equipment he would require to 
create this team. 'You can have whatever you need,' Watkins 
assured him. 

This was an offer that Altwasser could hardly refuse; on 
returning to Brussels he handed in his notice and prepared to start 
at Amstrad. During his interview he had told Watkins that he had 
worked with Sinclair from 1980 to 1982, but at the time this 
seemed of only passing interest. 

By the time Altwasser joined Amstrad in April, just as the deal 
with Sinclair was being completed, his previous experience could 
hardly have been more relevant. He was immediately thrown into 
collecting technical documents and circuit diagrams from people 
he had worked with in Cambridge. Amstrad drew on Altwasser's 
intimate knowledge of the Spectrum as it redesigned the Spec- 
trum's cabinet, built a cassette into the machine, improved the 



200 



ALAN sugar: the amstrad story 

sound, devised a new layout for the printed circuit board and 
modified the basic chip design. The company met its target of 
getting a Spectrum Plus Two into production by the end of July - 
time enough to produce the necessary volumes for Christmas 1986. 

Sugar set about making the Sinclair computer operations more 
efficient. 'There was not enough attention paid to quality control, 
stock control and monitoring of sales,' he said shortly after the 
acquisition. 

He saved money by dropping Sinclair's QL, the computer 
launched in 1984 which had turned into a quantum leap to 
nowhere. After placing an initial order with the Timex plant in 
Dundee, Amstrad also shifted production of the Spectrum to the 
Far East. Sugar had hinted during the press conference announc- 
ing the Sinclair purchase that the days of sourcing the Spectrum 
in Britain were numbered: 'We would like to manufacture in the 
United Kingdom, but we're a computer company, not a benevo- 
lent society.' 

In fact, Sinclair Research had already begun to source some 
Spectrums from South Korea before Amstrad took over. Looking 
back, Sinclair thinks the company probably should have switched 
manufacturing to the Far East earlier. 'Certainly if we had 
continued to produce a commodity product like the Spectrum, we 
would have had little choice but to produce in the Far East,' he 
says in retrospect. 

Originally Sugar thought that Amstrad might be able to come 
up with a radical new design for the Spectrum line; but Amstrad 
was constrained by one of the very factors which made the 
acquisition so attractive - the huge installed base of Sinclair 
machines. Amstrad had to ensure that the libraries of games 
already held by millions of youngsters could be used by Sinclair 
machines launched in the future. This need to maintain compati- 
bility severely limited the extent to which they could redesign the 
Sinclair range. 

Instead of working on a fundamental redesign of the Sinclair 
machines, Amstrad concentrated on improving what it had 
inherited. Altwasser found that the latest Spectrum embodied a 
third or fourth revision of the original chip design which he had 



201 



DAVID THOMAS 

worked on. The design had evolved haphazardly as logic chips, 
transistors and other components were added. Amstrad tidied up 
the interior, thereby reducing the manufacturing cost, and added 
some features, like a floppy disk version. But it stuck with the 
basic layout. 

In the event it hardly mattered since the Sinclair machine 
continued to sell well, particularly during the Christmas season, 
although the computer games market has declined from its peak 
in the early 1980s. In the second half of 1988, Amstrad was still 
selling £20 million worth of Sinclair machines, with sales almost 
equally divided between the U.K. and Spain, where Sinclair has 
long been a strong brand name. 

Sir Clive Sinclair, for one, believes that Sugar has done an 
excellent job in promoting and selling his old Spectrum range: 
The Spectrum was a commodity product. And I can't think of 
anyone who would have done as good a job as Alan Sugar did 
with it. I thought he was an ideal person to take over my Spectrum 
computer business. He knew what was needed by the customer. 
He was very, very perceptive. And he knew where to get the 
product made efficiently.' 

Amstrad can cope with the highly seasonal Christmas demand 
for the Spectrum precisely because it now has plenty of other 
products which sell all the year round. Alan Sugar explains: 
'Sinclair to me is like the suntan lotion that Boots sells. They put 
it in their windows in July and August and they take it away in 
the winter. And they can do that because they don't just rely on 
suntan oil. We can sell Sinclair because we can crank up produc- 
tion and stop it as soon as we want to.' 



As Richard Altwasser settled into his new job, he came to realize 
that Amstrad was defined by its laser-like concentration on 
products. It might be thought that all companies concentrate on 
their products, but Altwasser's exposure to a host of different 
concerns in his days as a management consultant had taught him 
otherwise. 

He explains: 'Some companies concentrate on the technology 
that goes into the product. Some companies concentrate on the 



202 



ALAN sugar: the amstrad story 

management structures of the business: their management gets 
into all sorts of knots about the administration of the business. It 
was clear that the thing that gave Alan and Bob a buzz was seeing 
new products brought out.' 

In particular, Altwasser learned to appreciate the difference 
between the Sugar and the Sinclair ways of working - a difference 
he summarizes as 'a concentration on products rather than a 
concentration on technologies'. 

At Sinclair, Altwasser says: 

There was a great deal of interest in the way in which materials 
and new technologies can make new components. The Spectrum 
used Ferranti ULAs [uncommitted logic arrays or gate arrays - 
a chip which is designed especially to satisfy the needs of 
particular customers]. These were high tech at the time and 
were the subject of great fascination to people in the company, 
Sir Clive included. But the detailed operation of the product, 
and how the user was going to perceive it, what features it was 
going to have, were far less interesting. I had a continual 
struggle to get Clive to sit down with the machine and show him 
what it could do and how it operated. 

At Amstrad, by contrast: 

I don't think I've ever discussed with Alan whether a particular 
chip uses 2 micron or \ x h micron CMOS, or whether it's two- 
layer metal or one-layer metal, or any of these other wonderful 
buzzwords. But I spend hours and hours discussing the features 
and functions, how they're presented and how the machine is 
going to be perceived by the customer. Alan doesn't get a buzz 
out of the technologies that go into making up the components. 
That's not something he finds exciting, but he's very interested 
in the features of the machine and how the machine operates. 

Sugar frequently wanders round the sixth and seventh floors of 
Amstrad's headquarters where Altwasser and his team of fifteen 
designers and engineers are based, but he is invariably there to 



203 



DAVID THOMAS 



talk about one of three things: features, reliability and cost. He 
remains totally obsessed with encouraging his designers to use one 
chip instead of two wherever possible: 'I like to visualize myself 
walking around with a pair of cutters. And when the engineer says 
he's got to have this, I look at it and I say, "How much is it?" And 
he says, "Five quid." So I say, "Cut it out." And he says, "It 
won't work." So I say, "Well, you have to find a way of making it 
work for a pound." ' 

Sinclair would not talk to his engineers in that vein. Indeed, he 
identifies himself as an engineer and not as a manager: 'I am an 
engineer, an inventor. I'm looking at very long-term projects - 
artificial intelligence or electric vehicles or things like that.' 

It is easy to exaggerate Sinclair's high-minded remoteness from 
the market. After all, he presided over the launch and marketing 
of two products - pocket calculators in the 1970s and home 
computers in the 1980s - which took the market by storm. His 
boffin image is partly a marketing tool: it reassured large numbers 
of technologically ignorant people who were buying such high tech 
products for the first time. 

Even allowing for the myths that were woven around Sinclair in 
the early 1980s, however, there is an important sense in which he 
and Alan Sugar are mirror images of each other. Physics journals 
jostled with books on micro-electronics and inventing on the desks 
in Sinclair's small offices in the West End of London at the end of 
the 1980s, reinforcing his image as the boffin par excellence. By 
contrast, Sugar's desk sat facing a television where he viewed the 
latest business news, while his office had a flip chart which he used 
to calculate the margins on his latest product. 

Sinclair seemed destined to be an inventor from the time he 
filled his bedroom with wires and amplifiers as a schoolboy, 
graduating to write manuals on transistors as a young man. His 
business roots lay in supplying the enthusiastic amateurs of the 
hobbyist market, his first commercial ventures involved selling kits 
to electrical buffs who wanted to assemble their own radios. 

Fascinated by the technology, Sinclair tended to neglect the 
business side of his ventures. Insufficient attention was paid to 
quality as some of his early products moved from development to 



204 



ALAN sugar: the amstrad story 

production, or while they were being manufactured. It was 
Sinclair's personal obsessions which drove the direction of his 
research rather than an appreciation of what consumers wanted. 
For every success, such as the calculator and the computer, there 
was a failure, like the mini-television and the electric car. Sinclair 
tended to become bored with his inventions precisely when they 
became successful, mass market products. He rapidly ceded his 
pre-eminent position to others - to the Japanese in the case of 
calculators, and to Sugar in home computers. 'Personally I don't 
like controlling a business that makes commodity products,' he 
acknowledges. 

Sugar is almost at the opposite extreme, going out of his way to 
emphasize his lack of interest in matters technological: 'We aren't 
here to get national awards for the greatest technology,' he said 
shortly before the Sinclair acquisition. The Amstrad boss takes 
great delight in affronting the sensibilities of the boffin class. Soon 
after the launch of his first word processor, he conceded that 
purists believed it to be as outdated as a crystal radio set. 'But it 
isn't. It's what the market wants. We've brought out a machine 
that operates on an elastic band. But it works,' he said, repeating 
the typical Sugar exaggeration which he had used in the privacy 
of Barclays Bank's headquarters during his negotiating sessions 
with the Sinclair team. 

In his heyday Sinclair was often praised by the Thatcher 
Government and by a gushing press as the very model of the high 
tech entrepreneur who would spearhead Britain's business revival. 
Burned by his sudden fall from grace, the opinion formers were 
wary of heaping the same expectations on Sugar. Yet it is arguable 
that Sugar, more than Sinclair, embodied precisely the qualities 
which British business has so notably lacked for much of the 
twentieth century. 

Britain is not short of boffins, clever scientists and technologists 
who devise new ideas; but it has lacked daring entrepreneurs 
prepared to translate those ideas into successful products, particu- 
larly products to compete in the high street with the offerings of 
the big battalions from Japan and the United States. There have 
been too few British businessmen with the determination to keep 



205 



DAVID THOMAS 

their eye on the ball, squeezing a product for every last penny of 
profit once it has become a success. Sugar has the confidence, the 
determination and the flair to do these things. 

At the first business meeting between Sugar and Sinclair in 
Liverpool Street station, the two men briefly toyed with the idea 
of joining forces. Sinclair would have invented the products and 
Sugar sold them. It could have been the perfect partnership, 'the 
dynamic duo', as Sugar jokingly called it. 

'I don't think it would have worked, because we're both too 
independently minded. I don't think our objectives are the same,' 
says Sinclair, looking back. He elaborates on this point: 'I think 
the difference really comes down to this: Alan makes products in 
order to make money, whereas I make money in order to make 
products.' 



206 



• 11 • 

Sticking to Its Knitting 



Alan Sugar was beginning to be noticed by the mid-1980s. After 
the word processor triumph, newspapers and magazines lined up 
to profile this rising star of the British business scene. He was 
talked about as a coming man, a representative of the rejuvenation 
of the British entrepreneurial spirit. 

Unknown to him, the Amstrad chairman was also being 
watched by one of the giants of the post-war industrial scene, 
Arnold Weinstock, head of the General Electric Company. Out of 
a series of brilliant mergers in the 1960s, Lord Weinstock had 
forged G.E.C. into Britain's biggest manufacturing company. He 
was a byword for managerial toughness, running G.E.C. through 
a mixture of rigorous financial controls administered from the 
centre and a well-developed system for devolving operational 
responsibility to G.E.C.'s constituent businesses. 

By the mid-1980s, Weinstock controlled his empire from Stan- 
hope Gate, a small road off Park Lane and one of the smartest 
addresses in London. G.E.C.'s head office is a tribute to Wein- 
stock's dislike of corporate extravagance. Its headquarters is a 
building so small and plain that it could be mistaken for an 
offshoot of a company a tenth its scale. Inside, Weinstock sits 
behind a large desk in the corner of a modestly sized room grilling 
his managers over the phone on the financial ratios of their 
businesses. The sole concession to luxury is half a dozen paintings 
of Weinstock's beloved racehorses; the only reading matter on his 



207 



DAVID THOMAS 

desk to lighten the wadges of financial data is a racing newspaper 
offering tips on the latest form. 

In the mid-1980s Weinstock came under fire for what appeared 
to outsiders as creeping stagnation within G.E.C. He seemed 
content to let the company tick over, sitting on G.E.C.'s cash pile 
of more than £1 billion. What few people realized at the time, 
however, was that he was already pushing ahead with the strategy 
which was to flourish in a string of major deals in the second half 
of the decade. G.E.C. was to spin off some of its most important 
subsidiaries into joint ventures with other large companies, keep- 
ing either controlling or large minority stakes in the resulting 
operations. 

G.E.C.'s managing director saw de-mergers of his businesses as 
an alternative to pressure from the City and the media for growth. 
He had come to the view that the commandment to grow 
continuously - which is fundamental to the way in which the City 
polices British business - was incoherent: 

Since the economy cannot grow at anything like the rate the 
G.E.C. has grown over the last twenty years, how is it possible 
to expect that we can just get larger and larger? Is that logical? 
I've always thought that can't be our ultimate fate -just to get 
larger and larger, and therefore more unmanageable. A more 
logical way of thinking is that some things become large enough 
not to be part of us. It's more than ten years since I started 
thinking about de-mergers. 

As Weinstock watched Sugar's progress, he began to think about 
Amstrad as a possible vehicle for de-merging some G.E.C. busi- 
nesses. Looking back, he describes why he was drawn to Sugar: 
'He has a talent which is not usually found in big electrical 
companies - his nose for the market. He has courage and nerve. 
He has commercial flair. He is young and independent. He has 
built up his business himself. I felt that we could harness that 
energy and those talents.' 

In those days G.E.C. and Amstrad had little in common beyond 
the fact that they shared Touche Ross as auditors. Weinstock 



208 



ALAN sugar: the amstrad story 

made discreet enquiries through the Touche Ross network about 
this relative newcomer to the British corporate scene, liked what 
he heard and so invited Sugar to Stanhope Gate. 

G.E.C.'s managing director and Amstrad's chairman hit it off 
at once. Weinstock, the son of a Jewish immigrant tailor, had 
begun his career in industry by working in the old Sobell television 
and radio company, later to merge with G.E.C. He charmed Sugar 
by reminiscing about the little Polish engineer who used to design 
television sets for him at Sobell, talking about those times as the 
good old days by comparison with the present when he was stuck 
in his office crunching numbers. The chat over, Weinstock got 
down to business and the two men discussed possible avenues for 
cooperation between their companies. 

After this initial meeting, Sugar phoned Howard Myles in the 
corporate finance department of his brokers, W. Greenwell. He 
explained to Myles the drift of his conversation with Weinstock 
and asked him to draw up a paper outlining options for partner- 
ship with G.E.C. Bound by Sugar to strict secrecy, Myles was 
forbidden to show the paper even to colleagues within Greenwell. 

Two main options emerged. The first would have meant G.E.C. 
spinning off its consumer products businesses into a company run 
by Amstrad. The second would have entailed G.E.C. taking over 
the whole of Amstrad and absorbing Sugar into the G.E.C. 
organization. 

It was the first option - a form of joint venture between Amstrad 
and G.E.C. - which received the more serious consideration. 
Amstrad would have taken over G.E.C.'s 'white goods' operations, 
one of the few parts of the company which manufactures products 
for sale in the high street. They were centred on Hotpoint, 
consistently one of the U.K. market leaders in products such as 
washing machines, refrigerators, tumble driers and dishwashers, 
but also included a number of other businesses. At the time of the 
talks, G.E.C.'s white goods business had sales of about £300 
million - similar to those of Amstrad - but its profits of £30 million 
were substantially below those of Sugar's company. In return for 
acquiring a package of G.E.C. assets, Amstrad would have issued 
new shares to G.E.C, giving it a minority stake in the enlarged 



209 



DAVID THOMAS 

Amstrad operation. Sugar would have run the new business, but 
he would have drawn on advice and other support from G.E.C. 
Weinstock explains how he saw the deal: 

Sugar seemed to me a way of de-merging our consumer products 
business. There was no reason why we should have taken him 
over unless that made managerial sense. There was more reason 
for him to run our consumer operations. We would have kept 
an interest. He could have been a part of G.E.C. but a part 
separated from the rest. He would have been the boss of what 
he was doing, but we would have been watching him and 
helping him. 

The joint venture held out other, less tangible benefits for G.E.C. 
which, unusually for an electronics company of its size, lacked a 
consumer electronics wing. It had held back from involvement 
with personal computers, while pulling out of television manufac- 
turing after an unsuccessful collaboration with Hitachi. Weinstock 
was aware that some of G.E.C. 's main rivals, such as Thomson of 
France, were immersing themselves more deeply in consumer 
electronics. Many people believed that new consumer electronics 
products, like high-definition television, would provide much of 
the growth in the electronics sector in the 1990s. Partnership with 
Amstrad would have furnished G.E.C. with a ready-made launch- 
ing pad into this business. 

The discussions about a G.E.C. -Amstrad partnership raised an 
even more delicate issue - Alan Sugar's standing as a possible 
successor to Weinstock. There was no obvious heir apparent 
within G.E.C, and Sugar would have brought a quality many 
observers believed lacking in Weinstock's empire: marketing flair 
to complement G.E.C.'s technical, financial and manufacturing 
strengths. 

Both Sugar and Weinstock were later to play down the sugges- 
tion that Sugar's suitability for the top G.E.C. post was a 
significant feature of the talks. Yet a deal between G.E.C. and 
Amstrad would have been interpreted by the outside world as a 
sign that Sugar was a candidate to succeed Weinstock. The option 



210 



ALAN sugar: the amstrad story 

of G.E.C. taking over Amstrad would have raised the issue most 
starkly: Sugar's position within G.E.C. would have had to be 
clarified. Yet the joint venture option also posed the question of 
Sugar's role, albeit in more submerged form, since the joint 
venture would have been seen as a half-way stage. Many outside 
commentators would have expected Sugar to be drawn further 
into G.E.C. if he made a success of running the company's white 
goods operations. 

Some of Weinstock's confidants expressed doubts. Within 
G.E.C.'s most senior councils, Sugar was described as 'strong 
medicine' — an entrepreneur who was used to running his own 
company in precisely the way he pleased. Amstrad's boss would 
have been coming into a large organization with established 
procedures and would have had to learn the importance of 
personal relationships in a big company - the necessity to deal 
tactfully with people who might be manifestly less competent than 
himself. In short, he would have had to learn patience, a quality 
which Weinstock himself had acquired only with some difficulty. 
Indeed, some said that two such strong-willed characters as Sugar 
and Weinstock would have found it hard to work together: even 
an organization as large as G.E.C. was big enough for only one of 
them. 

In the end, the Amstrad-G.E.C. talks foundered on more 
mundane matters. Initially both sides thought there was the 
making of a joint venture deal. Discussions advanced - though not 
to the stage of drawing up a draft agreement, because a price 
could not be agreed. Sugar thought he was being asked to pay too 
much for acquiring G.E.C.'s white goods operations, although the 
talks did not get as far as determining precisely which G.E.C. 
subsidiaries would be included in the package. 

As Sugar remembers it, G.E.C. wanted a premium price for its 
assets, as if it had been selling them in a normal commercial 
deal. But Sugar did not see it as a normal commercial deal; he 
was not happy even about basing the price on asset values, as he 
explains: 

G.EC wanted to value Hotpoint not only on the basis of its net 
asset value, but also as if we were buying it as a proper takeover. 



211 



DAVID THOMAS 



So there would have been a big premium. But I said to them, 
'All the Hotpoint plant and equipment is not worth two bob if 
we stop the production. I might go in there and say that I don't 
like the way you make motors. I might think I can go to Korea 
and get motors made at a quarter of the price, which will allow 
us to bump up Hotpoint's profits. But to do that I've got to shut 
the motor plant down and then all these so-called assets would 
be worth nothing.' 

Sugar's caution during the G.E.C. talks reflected a deeper reluc- 
tance on his part to become involved in corporate deals. He was 
suspicious of anything that would divert his energies from his 
prime task of developing Amstrad's products and selling them for 
all they were worth. In this, Amstrad was untypical of the 1980s. 
Sugar stayed clear of the City's game of buying and selling 
companies for ever more inflated prices. 

Many corporate stars of the 1980s grew big by gobbling up 
other companies. Until the stock market crash of October 1987, 
share prices appeared to be on a gravity-defying upward spiral. In 
this heady climate, companies were able to launch mega-bids 
based on paper offers by promising to pay for their victims with 
new shares issued at bull market prices. It was a game which kept 
hundreds of advisers in lucrative employment in the City, but it 
was not a game that Sugar chose to play. For a start, Amstrad's 
shares rarely traded at inflated prices; indeed, their value often 
struggled to keep pace with the company's spectacular perform- 
ance, partly due to the City's perennial nervousness about 
Amstrad. Moreover, Sugar could have entered the takeover game 
only on penalty of substantially diluting his controlling stake in 
Amstrad. 

Indeed, Sugar was reluctant to change any aspect of his strategy 
in order to please the City. In 1984 and 1985, for example, as 
Amstrad's profits mounted, his financial advisers badgered him to 
boost the amount of money paid out as dividends to Amstrad's 
shareholders. True, dividends had gone up, but nowhere nearly as 
rapidly as profits. W. Greenwell, Amstrad's brokers, thought more 
generous dividends would reassure the company's critics in the 



•212 



ALAN sugar: the amstrad story 

City. Howard Myles was invited to put Greenwell's case to an 
Amstrad board meeting. But Sugar argued strongly against allow- 
ing cash to seep out of the company in higher dividends, pouring 
scorn on the idea that City investors would begin to treat 
Amstrad's shares as a safe haven to be held through good times 
and bad if dividends were increased. Unsurprisingly, the Amstrad 
chairman won the day at the board meeting. It was a nice example 
of Sugar acting against his own narrow self-interest: holding 
almost 50 per cent of Amstrad's shares, he would have been by far 
the largest beneficiary of a more generous dividend policy. 

Sugar was not averse to acquisitions if they fitted clearly into 
Amstrad's existing business activities, however. The first such 
acquisition was of a minnow. In May 1984, just as it was entering 
the home computer business, Amstrad issued 250,000 shares (then 
worth about £240,000) to buy Lyons-Turner Enterprises, a whole- 
saler based near Stoke-on-Trent in the Midlands which specialized 
in distributing goods to independent electrical retailers. Barely 
noticed at the time, the acquisition was to bear important fruit in 
the evolution of Amstrad's computer strategy later in the decade. 

Four months later, in September 1984, Amstrad considered 
buying Fidelity, the hi-fi and television manufacturer. During the 
1970s, when the two firms had shared a sub-contractor in L. & N. 
of Rochester, Fidelity had been the larger company with the more 
powerful brand name. But it had lost its way in the early 1980s. 
The Citizen's Band radio episode was symptomatic of the differ- 
ences in fortunes: while Amstrad was nimbly entering and exiting 
the market, Fidelity was caught with £700,000 of unsaleable CB 
stock. 

On 4 September 1984, a £13.5 million takeover bid for Fidelity 
was unveiled by Caparo Industries. Caparo is an industrial 
holding company run by Swraj Paul, Indian by origin and one- 
time confidant of Indira Gandhi, India's late premier. Paul has 
moved easily between the business worlds of Britain and the 
Indian sub-continent ever since coming to London from Calcutta 
in the late 1960s. In 1984, he was keen to diversify out of Caparo's 
base in engineering, steel, tea and property into the higher tech 
market of consumer electronics. 



213 



DAVID THOMAS 

Sugar was interested in acquiring Fidelity's customer base and 
possibly also its factory in West London. A week after Caparo's 
bid, on 10 September, Amstrad announced that it had had 
informal contact with Fidelity and might launch a rival bid. 
Behind the scenes, however, Sugar was fully aware that Fidelity 
had run into problems with its production of televisions and 
cordless phones. Privately, a large institutional investor in 
Amstrad contacted Sugar to warn him to be wary: he had heard 
through the City grapevine that Fidelity was a can of worms. 
Three days later, on 13 September, Amstrad pulled out of the 
contest for Fidelity. Sugar explained that the latter's customer 
base was too similar to his own and that Amstrad's need for extra 
manufacturing capacity would be met by a further expansion of 
the Shoeburyness factory. 

With Amstrad bowing out, Caparo bought Fidelity later that 
month by raising its bid to £14.1 million, but Swraj Paul soon 
came to regret his victory. In July 1985, Caparo Industries issued 
writs against two former Fidelity directors and Touche Ross, its 
former auditors. Caparo was seeking damages to cover losses 
estimated at more than £10 million arising from alleged over- 
statement of profits and stock values at the time of the acquisition. 
Swraj Paul spent the following three years in a losing struggle to 
make a go of Fidelity. 

A similar episode occurred - although over an even shorter 
time-scale - in June 1987 when Thorn E.M.I., the large British 
conglomerate, sold its Ferguson television subsidiary to Thomson 
of France for £90 million. In the face of the onslaught from the 
giant Japanese television manufacturers, Thorn had decided to cut 
its links with Ferguson, the last substantial British-owned tele- 
vision maker. At the time of the sale, Ferguson was the largest 
player in the fragmented British television market, with a share of 
about 10 per cent. It had made almost 700,000 sets in 1986; it 
employed over 4,000 people, mainly at two factories in North 
London and Hampshire; it had a one-third stake in a joint venture 
with Thomson and Japan's JVC, making around 900,000 video 
recorders a year in plants in Britain, France and West Germany; 
and it had sales of about £300 million in the 1986-87 financial 



214 



ALAN sugar: the amstrad story 

year. In short, Ferguson was a substantial force in British con- 
sumer electronics. 

Thorn's sale of Ferguson to Thomson was announced on 
Thursday 18 June 1987. Alan Sugar was out of touch with the 
news for a crucial twenty-four hours and when he heard about the 
deal late on Friday evening, he was not amused. He immediately 
phoned the Financial Times to vent his displeasure. An FT reporter 
phoned him back, but communication was not at its smoothest 
because the Amstrad chairman was on a mobile phone which kept 
cutting out. By dint of repeated re-diallings and ten-second-long 
snatches of conversation, the reporter pieced together Sugar's 
concern: he was interested in both Ferguson's market share and 
its manufacturing plants. Although he had not yet studied the 
figures in detail, he would consider a counter bid. Amstrad could 
draw on its own cash pile to match Thomson's £90 million offer, 
unlike the French company which was having to borrow £40 
million from the banks for the purpose. Moreover, Amstrad would 
try to stop the deal by complaining to the European Commission 
that it was anti-competitive, because Thomson would totally 
dominate the European television market with Philips, the Dutch 
multinational. 

Sugar was angry that a British company had not been given the 
chance to bid for Ferguson. Describing the deal as 'a bit naff,' he 
told the FT, T just don't understand why they didn't approach 
us,' adding that Thorn had committed 'a dead diabolical liberty 
in not offering its Ferguson to us'. 

The Thorn-Thomson deal was in fact too advanced to be 
unwound. Sugar met Thorn executives the following week, but the 
company was determined to sell Ferguson to Thomson, its erst- 
while partner in the video recorder joint venture 'The time frame 
just didn't allow for us to interrupt. It was too far down the line,' 
Amstrad stated when it withdrew from the fray. 

Ferguson, Fidelity, Lyons-Turner Enterprises, together with the 
acquisition of Sinclair's computer business in 1986 - all fitted into 
a pattern. They were opportunities for Amstrad to build up its 
existing operations through buying market share, acquiring 
brands, adding to its manufacturing capacity or developing its 



215 



DAVID THOMAS 



distribution capabilities. None remotely followed the mid-1980s 
fashion for radical diversification. 

Amstrad was tempted to abandon this cautious attitude to 
diversification on only three occasions. The first was during the 
ultimately abortive negotiations with G.E.C. The second occurred 
when Sugar looked seriously at refrigerators and washing 
machines, an interest in white goods which had been sparked off 
by the G.E.C. talks. By May 1986 he was sufficiently confident to 
say publicly, Tt is absolutely firm that we will go into white goods 
- it is only a question of when. We are studying the market very 
carefully.' 

At first sight, white goods such as refrigerators and washing 
machines were an obvious business for Amstrad to contemplate. 
Their similarity to traditional Amstrad activities like hi-fi seemed 
close; both were consumer ranges sold in the high street and many 
of Amstrad's oldest customers, including Comet and Currys, were 
important white goods retailers. 

However, difficulties emerged as Sugar looked more closely at 
the white goods industry. One was precisely how Amstrad would 
enter the business. He had no intention of building his own 
manufacturing plant, not least because the sector was suffering 
from considerable over-capacity. The Far East was ruled out as a 
manufacturing base because it was not economic to ship bulky, 
low-margin products like fridges half-way round the world. The 
alternative seemed to be a link-up with an existing European 
manufacturer and Sugar talked about sub-contracting manufac- 
ture to an existing player. He considered buying a small producer 
in France, Spain or Italy, and in 1986 he had preliminary talks 
with Thorn E.M.I, about buying its domestic appliance businesses 
which included the Tricity, Bendix and Parkinson Cowan brands. 
Nothing gelled. 

Devising a marketing strategy for white goods posed an even 
more fundamental problem. Sugar's initial plan was to make such 
goods more like consumer electronics. Refrigerators and washing 
machines are not technology-drive, fashion-conscious items such 
as stereo sets or home computers; people buy fridges and washing 
machines out of necessity, not in order to own the latest model. 



216 



ALAN sugar: the amstrad story 

Sugar intended to change these entrenched buying patterns. 'The 
trouble is that at the moment nobody buys a fridge until the door 
falls off. The challenge is to turn it into something desirable, so 
that when the wife walks down the street she says, "I want that 
one,"' he said at the time. 

The strategy made sense; the difficulty lay in executing it. 
Amstrad hunted around for features which would transform the 
humble washing machine or fridge into a product which con- 
sumers would change every three years or so - without at the same 
time boosting prices beyond the reach of its traditional customers. 
This was to prove a fruitless search, as Sugar explains: 'When we 
realized that what we were talking about there were very, very 
bulky products at low margins, we decided we had no interest 
whatsoever. We could see no way of applying the Amstrad effect - 
of revolutionizing them, of putting features on them that would 
put us a cut above the rest of the industry.' 

The third time Amstrad was tempted into radical diversifica- 
tion was when it considered launching a hostile bid for Thorn 
E.M.I. One of the big four of British electronics, along with 
G.E.C., S.T.C. and Plessey, Thorn had been built up by an 
emigre businessman, Sir Jules Thorn, with many of the same 
swashbuckling entrepreneurial qualities as Sugar. But the com- 
pany had drifted after Thorn's departure, failing to make the 
transition from his individualistic and proprietorial style of 
management into a smooth-running, professionally organized 
multinational. In the first half of the 1980s a new chairman, 
Peter Laister, had evolved a vision of Thorn as an integrated 
communications and entertainment business. It would finance 
films and videos at one end of the production stream, make the 
televisions and video recorders on which they would appear and 
then distribute the materials through its own cinemas or shops. 
Pursuing vertical integration to its logical conclusion, Thorn even 
moved into chip manufacturing by buying the Inmos semicon- 
ductor company. Unfortunately, the grandiose vision was not 
matched by attention to details and Laister was abruptly 
removed by a board-room putsch in July 1985 after a sharp drop 
in the group's profits. 



217 



DAVID THOMAS 

After Laister's departure, Thorn was almost crying out for 
dismemberment; it had accumulated a ragbag of businesses which 
would be worth more on their own than as part of the sprawling 
Thorn empire. With Amstrad's share price riding high in 1986, 
Sugar cast an eye over the Thorn portfolio, asking his financial 
advisers to consider a bid. 'The idea at that stage was to break it 
up and keep the good bits,' he says. 

Amstrad's interest in Thorn, which was never publicly 
announced, did not get past the drawing board. Thorn seemed too 
big a morsel to swallow: its £3 billion turnover was then ten times 
that of Amstrad's. Tt was too much of a David and Goliath job,' 
Sugar says. 

In the event, Thorn's new top management began their own 
piecemeal dismemberment of the group, selling off parts peripheral 
to its new strategy including the Ferguson television business. 
With hindsight, Sugar is relieved that Amstrad did not push its 
interest in Thorn further: 

It would have been a terrible thing, because the problem with 
Thorn was bad management. And it was really like the pot 
calling the kettle black because we had no spare managers 
whatsoever. One of the things about big acquisitions is you've 
got to have spare brain power. You have to be able to say to 
your managers, 'Right, we've just acquired that big thing over 
there. You're in charge of it. Off you go, old boy.' We hadn't 
really got enough of those people at the time to look after our 
own business. 

Casting an eye over Thorn was one of several symptoms of a 
period of introspection within Amstrad in 1986 and early 1987. 
The company had managed a spectacular entry into computers by 
successfully launching its own products and buying the Sinclair 
brand. What would it do next? The link-up with G.E.C., the 
examination of white goods, a bid for Thorn - all reflected Sugar's 
restless search for the next great leap forward. To introduce order 
into this process, Amstrad called in the Boston Consulting Group, 
one of the world's most prestigious firms of management 



218 



ALAN sugar: the amstrad story 

consultants, to advise on future strategy. The kind of action which 
only a couple of years previously would have cut right across the 
grain of Amstrad culture, it was one of the first signs that Amstrad 
was aware of the need to evolve large-company procedures. 
However, the arrival of the management consultants coincided 
with the end of the period of uncertainty among their clients. 

Sugar cut short the consultants' stay in Brentwood, because 
they seemed to be echoing what he had already decided. Amstrad 
would not branch out on to some totally new path. It would stay 
with what it knew. It would grow organically, from within. It 
would stick to its knitting. 



219 



•12- 
Taking on Big Blue 



After the successful launch of the word processor in 1985, it 
became increasingly clear that Amstrad's future lay with com- 
puters. In the second half of 1985 Sugar mulled over what product 
should be introduced to follow the word processor. The options 
narrowed down to one: Amstrad would attack head-on the largest 
slice of the personal computer market - that dominated by I.B.M., 
the world's biggest computer company known throughout the 
business as 'Big Blue'. 

I.B.M. was slow to wake up to the personal computer. During the 
late 1970s, it watched from the sidelines as Apple and a host of other 
entrepreneurial companies pioneered the desktop revolution in the 
United States. Many of Big Blue's most powerful bureaucrats 
argued that I.B.M. should ignore this upstart industry in favour of 
what it knew best - selling its huge mainframe computers. By the 
beginning of the 1980s, however, even highly conservative I.B.M. 
had realized that personal computers were not a temporary fad. As 
the computing power which could be stacked onto desktops steadily 
increased, personal computers would inevitably become central to 
the computing strategy of all business users. 

In July 1980, I.B.M. surrendered to this infant technology by 
appointing an internal design team, which broke every rule in the 
I.B.M. book in order to meet its one-year deadline for creating the 
I.B.M. Personal Computer. Many of the parts for the machine 
were made by outside contractors rather than in I.B.M.'s own 



220 



ALAN sugar: the amstrad story 

factories; responsibility for the operating system was sub-con- 
tracted to Microsoft, a small software company led by a 25-year- 
old software wizard, Bill Gates. The MS-DOS operating system, 
written by Microsoft for the I.B.M. PC, was to make Gates a 
billionaire by the time he was 30. Belying its deepest instincts, 
I.B.M. gave its PC an 'open architecture'. Instead of keeping 
details of the circuitry and operating system secret, they were 
released to the world or, more precisely, to software companies 
which were encouraged to write business software for it. In that 
way, business users could be assured that they would find dozens 
of practical business applications to run on the I.B.M. PC. 

The I.B.M. Personal Computer, launched in August 1981, 
transformed the industry. The machine succeeded not because it 
was particularly innovative - many computer experts continued to 
prefer the rival offerings from Apple - but because it came with 
I.B.M.'s imprimatur. Before the company's entry into the market, 
many data processing managers would not have been seen dead 
with a personal computer; the new contraptions were fine for 
hobbyists, but were despised by the professionals who controlled 
the computer purchases of large corporations. I.B.M. changed all 
that. If Big Blue said personal computers were respectable, then 
they must be. The oldest cliche in the industry - that no one was 
ever fired for ordering I.B.M. - began to weave its magic. 

I.B.M. sold 156,000 personal computers in 1982, 455,000 in 
1983 and almost two million in 1984, according to market research 
specialists Dataquest. Four years after the launch of its PC, it was 
raking in annual revenues of $4 billion from its personal computer 
division - enough to have ranked the division as the 74th largest 
corporation in the United States if it had been a free-standing 
company. This phenomenal take-off exceeded the wildest hopes of 
the PC's champions within I.B.M. and, crucially, it also helped to 
introduce order into the personal computer market. The I.B.M. 
PC, together with its MS-DOS operating system, became the de 
facto standard in the industry. Apple was the only significant 
computer company to hold out against this encroaching 
uniformity. 

The computer industry seemed to be returning to a status quo 



221 



DAVID THOMAS 

which it understood: I.B.M. was king again. But this was to reckon 
without the birth of a new species of computer, the clone. The 
open architecture of the I.B.M. PC made it easy to copy. Elec- 
tronic entrepreneurs dotted around the United States began to 
design computers which were compatible with it but could be 
made in the Far East for a fraction of I.B.M. 's costs. A new clutch 
of computer companies was spawned on the back of these clones, 
which steadily eroded I.B.M.'s market share to the extent that in 
1985, sales of I.B.M. -compatibles overtook those of I.B.M.'s own 
personal computers. 

But the early clone wars were mainly fought in North America. 
I.B.M. did not launch its PC in Europe until early 1983, and at 
that time most of the successful U.S. clone companies lacked the 
resources or the expertise to tackle the European market. It was 
this vacuum in the European market - low-cost I.B.M. -compati- 
bles - which Amstrad set out to fill. 

Building an I.B.M. -compatible computer made sense to 
Amstrad for an additional reason. Soon after the launch of the 
PCW8256, it became clear that the word processor's overseas sales 
would not match those in Britain. The PCW8256 was perceived in 
the United States as a not particularly powerful CP/M-based 
personal computer, and Amstrad was too small a player there to 
educate the public into viewing it as a dedicated word processor. 
A similar fate befell the PCW8256 in West Germany, although 
there the fault lay partly with Amstrad's agents, Schneider, which 
promoted the machine as a personal computer and not a word 
processor, giving it a price tag to match. Unsurprisingly, the word 
processor did not take off in Germany as it had in Britain. 

Whatever the reason for the word processor's relative failure 
overseas, demands for a new flagship Amstrad product soon began 
to flood into the company's headquarters from its foreign oper- 
ations. Sugar recalls: 'We were mainly stimulated by our overseas 
dealers. They kept saying, "O.K., the CPC is very good, we have 
the Joyce, but now we want an I.B.M. -compatible." I said to 
them, "Sell the bloody word processor." And they replied, "No, 
no, no, we can't sell that. We want an I.B.M. -compatible."' 

In the early autumn of 1985 the specification was agreed for 



222 



ALAN sugar: the amstrad story 

Amstrad's I. B.M. -compatible computer, named the PC1512 - 
P(ersonal) C(omputer) and 512 after the memory size. Amstrad 
insiders took to calling the project the Airo (Amstrad's I. B.M. 
Rip-Off), though it was not a code-name that ever gained the 
currency which Joyce enjoyed. 

In fact, Amstrad went out of its way to ensure that its machine 
did not copy the I.B.M., since it knew how litigious Big Blue could 
be; it checked that the PC1512's keyboard layout and keyboard- 
computer connections differed from those on the I. B.M. MEJ 
Electronics was given the task of writing the BIOS - the basic 
input/output system, a key part of the operating system which 
controls communications between the central processor and 
peripherals such as disk drives and keyboard. Locomotive Soft- 
ware was originally asked to write the BIOS, but declined because 
it feared litigation from I. B.M. Mej located the BIOS in ROM 
(read only memory) - hence its alternative name of ROS (ROM 
operating system). Slipping easily into boffin language, Alan Sugar 
later described to PC User magazine the steps taken to ensure that 
Amstrad's ROS did not infringe I. B.M. patents: 

The ROS has been designed and engineered with no reference 
to the I. B.M. machine. It has been audited by lawyers in the 
U.S. and the U.K., and the writers of the ROS have been 
divorced from any I. B.M. ROS, and were just given a specifica- 
tion of what they had to do. We are quite happy that the 
independent auditors have compared our ROS with [the] 
I. B.M. [ROS] and that there is no similarity whatsoever. Also 
our cabinet is totally different. 

At the start of the project, Bob Watkins opened up an I. B.M. PC 
in order to look at the machine which Amstrad was planning to 
emulate. The technical director was amazed by what he found, as 
he recalls: 

I looked inside an I. B.M. PC and saw hundreds and hundreds 
of components. In our previous products, we'd designed our 
own gate arrays and I just couldn't understand why I. B.M. 



223 



DAVID THOMAS 

hadn't. There were these discrete devices all over the place and 
not one custom device in the whole bloody thing. It had massive 
boards stuffed full of chips. So we thought we can save a lot of 
money here by designing our own gate arrays. 

The design was constrained by the need to be compatible with 
I.B.M. technology. As a result, MEJ Electronics did not achieve 
quite the same level of integration as it had with the word 
processor. Nevertheless, it managed to design Amstrad's first 
I.B.M. -compatible using only three gate arrays, each replacing 
dozens of components in the I.B.M. computer. Amstrad was keen 
to repeat the word processor's winning formula of running as 
many features as possible off the computer's main printed circuit 
board, so Mej built into the machine many of the features that 
had to be bought as add-on expansion cards for the I.B.M. PC. 

Otherwise, Amstrad kept the cost low by its old tricks. It 
ordered parts in very large numbers; it priced its machine accord- 
ing to cost, not according to what the market would bear; and it 
took a cold look at what was really essential to a successful 
personal computer. Amstrad's computers, for example, were 
encased in plastic, unlike I.B.M.'s which came in metal. Sugar 
believes that a machine does not need to be gold-plated if it is 
designed for the mass market: 

There's two ways of designing a product. You either design for 
the military or you design for the mass market. Now a guy at 
I.B.M. would not think of using a keyboard with printed legends 
on the key caps. He wants the legends on the keycaps engraved 
in by a double moulding process. That means his keyboard is 
going to be £25, whereas ours will cost three quid. The difference 
between our keyboard and his is that in n years time the letters 
on our keyboard may rub off a little bit. But our philosophy is 
that after three years or so, you can buy another one which will 
be better and cheaper. 

Since Amstrad was entering the I.B.M. -compatible world, it 
needed standard software to run on the PC 1512. It tried to buy a 



224 



ALAN sugar: the amstrad story 

version of MS-DOS from Microsoft, but the American software 
company wanted to charge an un-Amstrad-like price for its best- 
selling operating system. So Amstrad turned to Digital Research, 
whose software had been closely associated with the company's 
CPC series and its word processor. 

Digital Research's Paul Bailey was shown a prototype of the 
PC 1512. He viewed it as the first real opportunity to create a 
mass market in business software in Europe, as he recalls: 'We 
were very excited about it. We thought that at the time it was 
extremely innovative to come in with a sub-£ 1,000 I.B.M.- 
compatible. Today, it looks routine, but in 1986 it was pretty 
revolutionary.' 

Bailey agreed to supply Digital Research's DOS-Plus operating 
system for Amstrad's personal computer. DOS-Plus was a belated 
attempt to catch up with the lead Microsoft had stolen by 
supplying MS-DOS to I.B.M. Essentially it was a clone of 
Microsoft's famous operating system, capable of emulating most - 
but, crucially, not quite all - of the capabilities of MS-DOS. 
Digital Research was happy to supply Amstrad with DOS-Plus at 
an aggressive price, because it saw Amstrad as one of its last 
chances to become established in the I.B.M. -compatible personal 
computer operating system market. 

Digital Research was also commissioned to adapt its Gem 
(Graphics Environmental Manager) software for the Amstrad 
personal computer. Gem allows many of the graphics-based 
features pioneered by Apple - like windows and icons - to be 
replicated on other computers. Gem also makes a computer look 
attractively easy to use when it is running in a shop window, so it 
was a natural addition to the Amstrad armoury. 

Digital Research was convinced that it could piggy-back huge 
sales of its software on the PC 15 12 and came out with a range of 
cheap business software programs adapted for Gem on the 
Amstrad machine. Bailey recalls that Digital Research was not 
disappointed: 'It was a big launch for the Gem concept in the 
U.K. We got behind the Amstrad launch in a very big way. It 
probably did more for Gem than almost any other initiative in 
those days.' 



225 



DAVID THOMAS 



Yet Amstrad still hankered after MS-DOS for its computer. 
Sugar realized that anything less than Microsoft's operating 
system would be unacceptable to many I. B.M. -compatible users, 
particularly in North America. Amstrad managed to set alarm 
bells ringing at Microsoft by convincing the American company 
that Amstrad's computer really was set to take the European 
market by storm. This was a sales opportunity which even the 
mighty Microsoft could not ignore. Shortly before the launch of 
the PC 1512, Microsoft dropped its price and Amstrad was able to 
announce that its machine would run both operating systems, MS- 
DOS and DOS-Plus. 

In early 1986 it was the open secret among computer industry 
watchers in Britain that Amstrad was working on an I.B.M.- 
compatible. The September launch took the razzmatazz of 
Amstrad's previous launches to new heights. A pair of giant red 
lips spelt out the range's specification on a huge electronic screen 
in front of 900 journalists, distributors, City folk and other pundits 
assembled in the Queen Elizabeth Conference Centre in central 
London. Asked to compute the cost of Amstrad's machine, the lips 
suggested a figure of £1,700, but added as an afterthought, 'Only 
one person in the galaxy knows the true price.' Enter Alan Sugar. 

Sugar unveiled eight machines in the PC 15 12 range. The least 
expensive, at £399 plus VAT, had 512k of memory and a single 
floppy disk drive. The most expensive, a colour machine with a 20 
megabyte hard disk, cost £949 plus VAT. The cheapest I. B.M. 
personal computer at the time sold to corporate clients in Europe 
for £1,429, although one computer reviewer argued that the 
nearest equivalent I. B.M. cost over £4,000. 

Amstrad's chairman was in confident mood at the launch and, 
in a bid to generate momentum for his new computer among 
retailers and dealers, he released bullish sales forecasts - forecasts 
which were to return to haunt him. He stated that production of 
the new machines in the Far East would build up to 70,000 a 
month almost immediately, and could increase to 100,000 a month 
early in 1987. He aimed to sell 300,000 by the end of 1986 and 
800,000 in a full year, over half outside Britain. 

Many specialist commentators contributed to the euphoria. 



226 



ALAN sugar: the amstrad story 

Computer journalist John Lettice, writing in the Observer, said that 
the Amstrad stood out from the crowd not just because of its price, 
but also because of what users would get for their money: free 
software, built-in devices for communicating with printers and 
other computers, a mouse pointing device and so on. The Guardi- 
an's reviewer, noting that Amstrad's Intel 8086 microprocessor 
made it faster than the I.B.M. PC, concluded, 'While Amstrad is 
famous for cost cutting, the company has not skimped on the 
PC1512's performance.' Guy Kewney's five-page bench-test for 
Personal Computer World was ecstatic: 'It goes faster than the I.B.M., 
it's smaller, has better onscreen colours, and includes functions 
which have to be added (and paid for) separately on normal 
I.B.M. -style machines.' Which? consumer magazine confirmed 
these specialist judgements in May 1987 by naming the PC1512 
its personal computer best buy. 

Amstrad backed the PC 1 5 1 2 with a massive and typically aggress- 
ive advertising campaign. 'Compatible with you know who. Priced 
as only we know how' was the slogan devised by Amstrad and 
Delaney Fletcher Delaney for the PC 1 5 1 2. The ads showed a whole 
Amstrad as being equivalent to half an I.B.M., and even less of an 
Olivetti and Compaq. Fearing trouble from I.T.C.A., the television 
advertising watchdogs, Peter Horrell - Amstrad account director at 
the advertising agency - bought I.B.M., Olivetti and Compaq 
machines and used a laser cutter to slice them in precise ratios 
according to price. 'That ad took a lot of pre-planning to set up. We 
had to demonstrate that we were comparing like with like in terms of 
the machines' specifications. If anything, we were over-generous to 
the competition,' Horrell recalls. 

Yet the PC 1512 soon ran into a barrage of bad publicity which 
dogged the computer for the rest of 1986. The first and most 
serious allegation was that the PC 1512 overheated because it 
lacked a fan. In mid-October, the Sunday Times published an article 
reporting that chemicals giant I.C.I, had decided against buying 
the Amstrad computer. The Sunday Times attributed this 'decision' 
to fears that the PC 15 12 could prove unreliable when used 
continuously on a computer network. It also quoted Mike Shin- 
gler, supposedly an adviser to I.C.I, on computer purchases, as 



227 



DAVID THOMAS 

saying that lack of space and ventilation inside the Amstrad 
cabinet could cause overheating. 

The Sunday Times report was followed by a deluge of rumours in 
the computer trade press about overheating and other problems 
with the PC 15 12. Part of the difficulty seemed to stem from the 
fact that, unlike other computer manufacturers, Amstrad had 
placed its power supply in the monitor, thereby removing any 
need for a fan in the computer's system unit. Yet the more 
penetrating computer journalists also detected something amiss 
with these stories. Guy Kewney described how PC Dealer had been 
bombarded with tips about problems with the Amstrad from 
hushed informers: a leading distributor had returned its first batch 
of 1512s because they were all faulty, 'the keyboard will wear out 
within three months', and so on. On checking, these tips turned 
out to be bogus. The computer magazine Microscope blew the gaffe 
in late October, when it reported that individual I.B.M. salesmen 
had been spreading rumours among computer dealers about 
overheating Amstrads. 'We were told that I.C.I, tested out the 
Amstrad with a token ring (networking) card and that the casing 
started to melt,' one I.B.M. dealer told Microscope, 

Alan Sugar was in understandably angry mood when he flew 
back from the Far East in the last week of October to confront 
these rumours. He compared the behaviour of salesmen working 
for some large personal computer companies unfavourably with 
his experience at the cut-throat end of the audio market: 'When I 
was competing against 40 other small-time dealers, similar to 
myself, who'd kill their grandmothers in order to beat me to a deal 
. . . these dirty tricks and lies were never thrown at us.' 

The Amstrad boss was particularly incensed because the relia- 
bility of the CPC range and the PCW8256 had been important to 
their success: fault rates for the CPC464, Amstrad's first computer, 
had been notably low in the days when home computer manufac- 
turers were a byword for poor quality. Sugar was affronted at the 
implication that Amstrad had failed to carry out elementary 
quality testing of its machines: 'We've been in the business, in hi- 
fi, for twenty years. We have environmental chambers. We stimu- 

•228- 



ALAN sugar: the amstrad story 

late the worst conditions in the field - drop tests, heating tests, 
pissing-on-it tests, you name it,' he told PC Dealer, 

Indeed, the original Sunday Times report was quickly shown to 
be wide of the truth when I.C.I, stated that it was in the middle of 
lengthy tests of the PC 151 2, and that the executive quoted in the 
article had no authority to speak on its behalf on this subject. In 
January 1987, after completing the tests, I.C.I, wrote an open 
letter to Amstrad which said: 'During the trials no problems were 
experienced with overheating when the Amstrad PC was con- 
nected to a token ring network. The Amstrad PC 1512 has now 
been approved for purchase by I.C.I, operating units.' 

Another charge thrown at Amstrad was that it was almost 
impossible to buy one of the now famous PC 15 12s. This complaint 
was nearer the mark for initial demand exceeded Amstrad's most 
optimistic projections. Indeed, in late October Amstrad felt confi- 
dent enough to announce a 12.5 per cent price increase on the 
PC 1512, blaming this on the rising value of the yen - the currency 
in which many of its supplies were purchased. One dealer told 
Microscope magazine of people placing orders for the Amstrad over 
the phone without bothering to look at the machine. 'I'd say it's 
cut demand for the Olivetti by at least 50 per cent. Apart from 
anything else, the poor bastards who want Olivettis can't get 
through because the lines are jammed by people calling about the 
Amstrad,' he said. 

Just as important, the pattern of demand differed from 
Amstrad's expectations. Since its computers were so cheap, many 
customers ordered the top-of-the-range, hard disk machines. 
Amstrad had expected hard disk sales to account for up to 10 per 
cent of demand, but they were running initially at nearer 35 per 
cent. Sugar made a rapid sweep through Taiwan, Singapore and 
Japan in the second half of October in a vain bid to find additional 
supplies of hard disks. 

In early November Sugar took the opportunity of an interview 
with the Financial Times to announce how he planned to deal with 
both the hard disk and the overheating problems. Amstrad would 
be shifting the mix of its range towards the more expensive, hard 
disk machines; this shift was possible because it had placed orders 



229 



DAVID THOMAS 

for the disks with Tandon and for the control mechanisms with 
Western Digital, two suppliers based in the United States. 
Although still vigorously denying that his computers overheated, 
Sugar had decided to put a fan into all Amstrad's hard disk 
machines in a bid to kill the overheating rumours. In what was to 
become his classic statement of the customer-is-always-right prin- 
ciple, the Amstrad chairman told the startled journalist: 

I'm a realistic person and we are a marketing organization, so if 
it's the difference between people buying the machine or not, 
I'll stick a bloody fan in it. 

And if they say they want bright pink spots on it, I'll do that 
too. 

What is the use of me banging my head against a brick wall 
and saying, 'You don't need the damn fan, sunshine'? 

Shortly after these comments were published, pink spots appeared 
on computers in the most surprising places, reportedly even in 
Amstrad's headquarters. The staff of one computer magazine 
came in to work to find its computers covered in pink spots, 
together with a little label: 'Congratulations, your computer has 
been upgraded to Amstrad compatibility.' 

Sugar's swift reaction dampened down the rumours, but did not 
manage to kill them entirely. Questions about the PC1512's 
reliability continued to appear in print into 1987, and were 
immediately met by threats of legal action from Amstrad. Libel 
writs from Amstrad forced several prominent organizations to 
make humiliating retractions of statements made in their name 
about Amstrad's computer, including the B.B.C. which apologized 
unreservedly for an article in its staff magazine which threw doubt 
on the PC 1512's safety. 

In the autumn of 1986 some pundits spotted a more fundamen- 
tal problem in Amstrad's computer strategy, a problem which at 
first went almost unnoticed amidst all the froth about overheating 
machines. The PC 1512 did not sell to large corporate customers 
as well as Amstrad had hoped. 'How well are your business sales 
going?' was later to become a query with which Alan Sugar was 



230- 



ALAN sugar: the amstrad story 

routinely confronted by journalists and City analysts. It was a 
question which irritated him because, as he pointed out, most of 
Amstrad's PC range was sold to people with VAT registrations - 
small business people or professionals working from home. Yet his 
answer could not entirely disguise the fact that Amstrad had 
expected to sell larger quantities to big corporations. 

Some cautious computer buyers in the corporate sector had no 
doubt been frightened by the overheating publicity, while the more 
feeble were perhaps put off by Amstrad's lingering image as a low 
end audio company. The more technically minded pointed to the 
PC1512's lack of an EGA (Enhanced Graphics Adaptor) capabil- 
ity as a reason for not buying it. Soon after the launch of the 
PC 15 12, EGA became the industry-standard system for control- 
ling the colour resolution on a personal computer screen. But the 
most important reason for Amstrad's difficulties in penetrating the 
corporate market was undoubtedly the way it actually sold its 
computers. 

Amstrad's personal computers continued to shift in greatest 
numbers through chains like Dixons or Wilding, which served 
individual users or small businesses. By contrast, corporate pur- 
chasers were accustomed to buying either from specialist computer 
dealers or direct from a computer manufacturer's own sales force. 
Like some of its rivals, Amstrad did not have its own direct sales 
force, relying instead on specialist dealers for sales to professional 
buyers. Unfortunately, the inevitable corollary of Amstrad's low 
prices was tight margins for dealers: one estimate suggested that 
dealers made profits of about 12 per cent selling Amstrad's 
computers, as compared with approximately 30-40 per cent on 
rival machines. Crucially, the margins available to dealers on 
Amstrad computers made it difficult for them to offer corporate 
customers back-up consulting or repair services. Anticipating this 
problem, Amstrad arranged for a service company, Dictaphone, 
to offer maintenance contracts to users of Amstrad computers. Yet 
the fact remained that a package which was fine for small 
businesses and individual users - low prices, but minimal service 
- transferred awkwardly into the large corporate market where 
spending money for a hassle-free life was the norm. This was an 

•231 • 



DAVID THOMAS 

issue which was to assume increasing importance as Amstrad's 
computers moved up-market. 

Amstrad's critics dived for cover in February 1987 when the 
company announced its results for the six months to the end of 
December 1986 - reflecting the first burst of PC 151 2 sales. Profits 
advanced 159 per cent to £71.3 million, on sales up 112 per cent 
at £272.5 million. Amstrad shares closed 19p up on the day at 
166p, adding £47 million to the value of Sugar's personal stake 
in the company. A cautious Lex note in the Financial Times 
pointed out that the company had far to fall if its 'prodigious 
run of marketing triumphs' was to end: only about 15 per cent 
of the £900 million price tag the City was then putting on 
Amstrad reflected tangible assets - the rest reflected 'goodwill', 
essentially the Amstrad brand name and the market's confidence 
in Sugar. 'The depressing thought is that ever bigger marketing 
coups are required just to keep up momentum,' Lex commented, 
before adding in an unusual burst of adulation, 'The less 
depressing thought is that the company is run by a marketer of 
genius.' 

For the moment, Amstrad was riding high in City eyes. About 
50 per cent was added to the company's share price between 
Christmas 1986 and February 1987, as investors heard the early 
news about PC 1512 sales. Romtec, a microcomputer market 
research organization, reported that the PC 15 12 had shot to the 
top of Britain's personal computer market in December, the first 
month in which the machines were widely available. Amstrad 
accounted for 26 per cent of the personal computers sold through 
dealers in December, as compared with 25 per cent for I.B.M., the 
traditional market leader. This lead was even more substantial 
once allowance was made for high street sales, although the higher- 
priced I.B.M. was still comfortably ahead in terms of sales by 
value. Romtec confirmed that the PC 1512 had achieved the 
'Amstrad effect': Amstrad was expanding the overall market rather 
than eating into its competitors' sales, which in most cases had 
been largely unaffected. Despite the lower margins, dealers were 
bowing to the inevitable - selling Amstrads to customers who 
simply would not have bought a more costly machine. 



232 



ALAN sugar: the amstrad story 

Yet market reports by independent observers such as Romtec 
held a double-edged message for Amstrad as 1987 unfolded; they 
continued to confirm that Amstrad was leading the pack in Britain, 
but they also suggested that the company would not hit the 
ambitious sales targets announced by Sugar when launching the 
PC 1512. The story was similar on the Continent: Amstrad was 
selling large volumes, but not large enough to reach its targets. It 
began to appear that Amstrad might be selling 'only' about 40,000 
PC 1512s a month - far short of the minimum of 70,000 a month 
predicted at the time of the launch. 

By all normal standards this was still a phenomenal success, 
considering that Amstrad was new to the I. B.M. -compatible 
market. The trouble was that City analysts had based their 
forecasts on Sugar's launch volumes; they had heard the Amstrad 
chairman's predictions, estimated the margins, picked up their 
mobile phones and told their firms to buy Amstrad stock for all 
they were worth. When the truth began to emerge from Romtec 
and similar organizations, they hastily downgraded their forecasts, 
producing a sharp drop in Amstrad's share price. In the first three 
weeks of June 1987, Amstrad stock fell some 50p, knocking almost 
£300 million off the company's market value. 

Relations between Amstrad and the City reached a new low 
when Chase Manhattan Securities slashed its profit forecast in 
early June. Chase had been one of the most bullish followers of 
Amstrad, talking up its profit forecast despite discreet warnings 
from Amstrad headquarters that it was becoming too 
enthusiastic. 

Two events prompted Keith Woolcock, a young analyst with a 
colourful turn of phrase who was then at Chase, to drop his 
optimistic forecasts for Amstrad. First, a long phone conversation 
with a senior Amstrad executive convinced him that the company 
was selling nothing like 70,000 PC 1512s a month. Second, he read 
a report in the Financial Times that Amstrad had just launched a 
new and more powerful personal computer, the PC 1640, in the 
United States at the Comdex computer show in Atlanta, Georgia. 
The story quoted Alan Sugar as saying that the computer would 
not be launched in Europe until early in 1988. Woolcock recalls 



233 



DAVID THOMAS 



his reaction: T thought: "This is a heap of crap. He's lying." I 
reckoned he was going to launch the PC 1640 in Britain too, and 
the only reason for doing so was because the PC 1512 was not 
selling as well as expected.' 

Immediately he read the FT article Woolcock changed his mind 
about Amstrad's prospects. 'I had one of my Pauline experiences. 
I stood up at our morning meeting and said, without a note, I'm 
cutting our forecast.' 

No company welcomes a profits downgrading from a big City 
institution, but Woolcock rubbed salt in the wounds by publicly 
urging Amstrad to come clean with the City. He was quoted in 
the Sunday Times as likening Sugar to a Trappist monk who refuses 
to talk to anyone except to complain about his critics. On Channel 
4 television, the young analyst urged Sugar to emulate the prophet 
Mohammed by coming down from his mountain to talk to the 
masses. Now Woolcock muses: 'Being a Jew, that probably didn't 
go down too well. He'd hit the City pretty hard and I thought he 
wouldn't mind some similar treatment. Alan Sugar isn't as thick- 
skinned as you'd think.' 

The thickness of Sugar's skin became abundantly clear a week 
later when a letter arrived from Amstrad's solicitors threatening 
to sue Woolcock and Chase - a move believed to be without 
precedent even in the often fraught history of relations between 
quoted companies and the stockbroking fraternity. The letter 
accused Woolcock of libel and said that if Chase were to 'lie about 
Amstrad, or its associate companies, managers, products, perform- 
ance or sales figures in future, Amstrad will enforce its rights 
without notice. This is both a threat and a promise.' 

More seriously still, Amstrad suggested that Chase had tried to 
manipulate the Amstrad share price. 'Sugar accused me of acting 
like a spiv. He said I'd had a profit forecast wliich was much too 
high and then cut it, having first squared our market makers book 
so we could make a tidy profit on it,' explains Woolcock, who 
insists that on the contrary Chase lost money as a result of his 
forecasts. 

Chase's lawyers wrote back a terse letter rejecting the charges 
and Amstrad did not press the matter, but diplomatic relations 



■234 



ALAN sugar: the amstrad story 

with Chase were ruptured for some months. Woolcock next saw 
Sugar in November at Amstrad's annual general meeting. Held at 
the Waldorf hotel in central London, the gathering seemed to the 
young analyst to be full of private investors who worshipped the 
ground that Sugar stood on. ('I felt [as if] I was running against 
the bulls in Pamplona,' Woolcock recalls.) He sat next to a middle- 
aged woman in a blood-red dress, dripping with jewellery, who 
told Woolcock that Sugar had changed her life: she had bought 
her first shares after seeing him on television and now she was an 
avid stock market follower. Looking around, she said, 'Alan's done 
the same for a lot of people in this room.' ('Screw British Telecom: 
this really is popular capitalism,' thought Woolcock.) 

When the meeting started, Sugar was asked why Amstrad's 
dividends had not increased as rapidly as its profits. The chairman 
leapt at the question as though it were sent from heaven. If he 
increased the dividend, he said, it would just attract the kind of 
institutional investors who ran for cover every time they received 
an ignorant note from some analyst. The private investors cheered, 
but Woolcock was beginning to feel distinctly nervous: 'I was 
shitting myself, expecting that any minute Sugar would point to 
me and say: "And that's the man who's been telling people to sell 
your shares."' 

The year 1987 did nothing to improve Sugar's already low 
opinion of the City. 'There should be some professional exam for 
these analysts. Most of the time they talk through their backsides,' 
he said at the time. The scare stories about the PC1512's perform- 
ance also reinforced his contempt for much of the specialist 
computer trade press. A chill descended on relations between them 
and Amstrad, which at best had been never more than lukewarm. 
For a time, many journalists were lucky to find anyone at Amstrad 
prepared to talk to them. 

Why should Amstrad care? As far as Sugar was concerned, 
more reliable measures of his success in introducing the PC 151 2 
could be found in Amstrad's results for the financial year 1986-87, 
released at the end of September 1987. Profits were up 80 per cent 
at £135.7 million; sales were up 68 per cent at £51 1.8 million; and 



235 



DAVID THOMAS 



profits as a percentage of sales were at an all-time Amstrad record 
of 26.5 per cent. 

To underline the spread of its interests, Amstrad broke its 
results into two main groupings: leisure, including audio, video 
and home computers such as the Sinclair and CPC ranges; and 
business, including its word processors and I. B.M. -compatibles. 
Business products accounted for 54 per cent of sales in 1986-87, a 
sharp increase on the 38 per cent of turnover they had contributed 
the previous year. Another year and another metamorphosis in 
Amstrad's many lives: for the first time the company was making 
most of its money by selling equipment to small businesses and 
other professional users. 

Just three years after launching its first computer, Amstrad had 
established four broad ranges: Sinclair, CPC, word processor and 
I. B.M. -compatible. Upgrades could now be slotted into these 
ranges without any fuss: the PC 1640, with better graphics capabil- 
ity, was unveiled in June 1987, followed in November by the 
PCW9512, with its letter-quality printer and an improved version 
of the Locoscript word processing program. 

After a year of buffeting from the media and the City, Sugar felt 
justified in a little crowing. The PC 1512, he wrote in his chair- 
man's statement, 

... has revolutionized the so-called computer industry in the 
U.K. The 'Amstrad effect' of low price high volume production 
and marketing has taken the wind out of an industry cocooned 
in a sedate and uncompetitive environment, encapsulated in a 
hyped-up aura of mystical high technological sophistication. 
The 'Amstrad effect' of course rocked the boat, the resultant 
factor being critical comments of the product with which I am 
sure all and sundry are familiar. I think Pythagoras and 
Columbus had the same problem when they announced the 
world was round. 

The heralding of yet another phase in Amstrad's history was the 
most striking feature of Sugar's statement. It was unrealistic, he 
wrote, to expect Amstrad to continue doubling in size every year. 



236 



ALAN sugar: the amstrad story 

'Clearly a new strategy is necessary,' he explained. 'The current 
financial year is going to be one of consolidation and seed planting 
in formulating our firm plan of growth for the future. A major 
stepping-stone for Amstrad and one which, I predict, in years to 
come will be recognized as one of the major strategic milestones in 
the company's history.' 

Alan Sugar was about to usher in another new Amstrad. 



237 



• 13- 
Amstrad Culture 



The Amstrad which launched its first I. B.M. -compatible computer 
in 1986 was still run like a small company. Alan Sugar's style of 
managing Amstrad was little changed from his early days in the 
audio market. True, he had needed to delegate more as Amstrad 
grew, and by 1986 day-to-day relations with even some of the 
largest customers were handled by his subordinates. But to carry 
out his wishes, Sugar had gathered round him a group of core 
workers similar to himself. They were sharp, aggressive, down to 
earth, pragmatic and convinced that Amstrad deserved to be on 
top of the pile. Amstrad had none of the diffidence to be found in 
most British organizations; in this respect, it was more like an 
American corporation. 

If you wanted to work at Amstrad, you had to be ready to take on 
the world. One visitor to its Brentwood headquarters in Essex in 
1986 came away likening the company work-force to a troop of 
Israeli paratroopers. If Amstrad was an elite force of shock troops, 
then the identity of the general was never in any doubt. Alan Sugar 
directed his team from a battered leather armchair in the middle of 
an open-plan office at the top of the building, on the ninth floor. He 
might be a multi-millionaire; he might be widely profiled as Britain's 
new business phenomenon; but that was no reason for him to indulge 
in the luxury of his own office. 

This lack of pretension was a feature of the Brentwood head- 
quarters. In 1986, Amstrad's products were arranged haphazardly 



238 



ALAN sugar: the amstrad story 

on the shelves in the company's boardroom. Smudged fingerprints 
competed for attention with sales and production charts on the 
bare white walls of the offices. Foam rubber poked through the 
seams of the chairs in a conference room, a journalist from Time 
magazine reported just after the launch of the PC 15 12. 

The Amstrad culture attracted unconventional people and only 
tough, unconventional people survived. The story of one young 
man who joined just before the launch of the PC 1512 may be 
extreme, but it gives a representative flavour of what it was like to 
work there. 

In 1986, Thomas Power was a fresh-faced 22-year-old, working 
in a small advertising agency in Croydon, south London; this was 
his first job since finishing a course in marketing at a local college. 
Ambitious and restless, he was looking around for a better 
opportunity and had begun to read in newspapers and marketing 
magazines about the exploits of Amstrad, a company by then 
being described as the marketing success of the decade. In a 
moment of candour, Power confided in his father that his goal was 
to work for Amstrad. 

In late May, Power noticed an advertisement for a marketing 
manager at Amstrad, precisely the job he wanted. Unfortunately, 
it was clear from the advert that Amstrad was looking for someone 
in their late twenties or early thirties. Undaunted, he persuaded 
two of his friends to combine their CVs with his and submit a joint 
application for the job. This unconventional approach must have 
intrigued Malcolm Miller, because the marketing director called 
them in for an interview. The three young men put on a presenta- 
tion, complete with flip charts, explaining to him why they were 
the triumvirate that Amstrad badly needed. 

One of Power's friends dropped out, but he and his remaining 
partner were interviewed four times over the next five months. At 
the end of this assault course, Amstrad offered them both jobs, 
even though Miller was clearly more impressed with Power's 
friend. The latter then turned Amstrad down, whereupon the job 
offer to Power was withdrawn. But Power had become obsessed 
with working for Amstrad and early one morning he travelled to 
Amstrad's headquarters, buttonholed Miller as he arrived for work 



239 



DAVID THOMAS 



and asked for another chance. He handed Miller a letter explaining 
why he - alone of the three friends who had originally applied for 
the job - deserved to work for Amstrad. This apparently ruthlessly 
back-stabbing letter had in fact been drafted by Power and his two 
friends the night before. Miller looked at the letter and said, 
'You've got three months.' 

Power floated out of Amstrad's headquarters, having achieved 
his ambition. The last thing he expected was that the next six 
months would be, in his words, 'the hardest punishment I've ever 
been through in my life'. 

Amstrad's new marketing manager was in for a shock when he 
arrived at work bright-eyed and bushy-tailed in a new suit in 
October 1986, one month after the launch of the PC 151 2. His first 
surprise was to realize that Amstrad, then a company turning over 
more than £300 million a year, had no marketing department. 
Before his arrival, Sugar and Miller had handled all marketing 
matters; henceforth Power was the marketing department. His 
second shock was to find everyone too busy to tell him what to do; 
he was given no training or even introduction to the way Amstrad 
did business. The only guidance Power had was his job description 
- a piece of paper with ten points on it, including product 
management, media buying and advertising. 

When Power arrived on Monday morning, Malcolm Miller 
simply told him to get on with it. Shortly afterwards, Miller 
introduced the new man to Amstrad's chairman: 

Miller: Thomas, this is Alan Sugar.' 

Power: 'Hallo.' 

Sugar: 'What are you doing here?' 

Power: 'I've come here to do marketing.' 

'Huh,' grunted Sugar walking away, as if to say: 'marketing - 
what do you know about marketing?' 

Power was given a little table at the far corner of the big open- 
plan office on the ninth floor - the room dominated by the large 
desk and battered armchair occupied by Sugar. Grouped around 
Sugar's desk were some smaller tables at which half-a-dozen old 
Amstrad hands sat. 

When the working day started, Power was in for his third and 



240 



ALAN sugar: the amstrad story 

biggest shock, as he recalls: 'In the middle of this room was this 
man shouting at the people at all the tables round him. Alan was 
literally standing in the middle of these tables, barking orders at 
people. And every now and then he would stare at me out of the 
corner of his eye as if to say, "What on earth can you contribute 
to this organization?" And I would stare back and think, "Yes, 
what can I contribute?"' 

Once Sugar had given his orders to the men at the tables, they 
would pick up their phones and shout instructions down the line 
to Hong Kong, Japan or wherever. Sugar surrounded himself 
with the people immediately involved in shifting Amstrad's 
products. They were directing sales processing, exports, shipping 
and distribution. The managers responsible for manufacturing, 
finance or product development were elsewhere in the building. 
Three questions dominated everything on the ninth floor: How 
many have been sold? How many are in stock? How many are in 
shipment? 

Sales were at the heart of the Amstrad empire and the nearer 
you approached the flame at its heart, the more likely you were to 
get burnt. On Amstrad's ninth floor Thomas Power was close to 
this flame. The ninth floor induced severe cultural shock trauma 
in the young man. He had left the glitzy world of advertising - one 
of tasteful furniture, designer carpets and pretty secretaries. And 
for what? T seemed to be working with a bunch of football 
hooligans who were trading. It didn't matter that they were 
trading hi-fi and computers. They could just have easily been 
trading in bananas, orange juice, coffee, bacon - it would have 
made no odds. We just had to ship this stuff and sell it as fast as 
possible. The only atmosphere I know that is similar is a dealing 
room in the City,' Power comments. 

The pack turned on Power in the early days when he was 
referred to derisively as 'the marketing man'. Everyone acted as 
though he would not last long. During his stint at Amstrad, 
he saw many young people arrive only to disappear abruptly 
after two or three weeks. There was no ceremony in their 
leaving. 

But Power survived and carved a job out of what he knew - 



241 



DAVID THOMAS 

advertising, media buying, product launches. Gradually he 
learned about sales forecasting, the seasonal fluctuations of the 
products, the distribution channels, differences between overseas 
markets and so on. He learned also about his powers of endurance. 
'You cannot work for Alan Sugar unless you are of sound mind. If 
you are one of these people who takes things personally, you will 
have no future with that company. You have to let it slip off your 
back,' he explains. 

Power needed the support of his father, to whom he was 
unusually close, in order to last his first three months at Amstrad. 
Every weekend he drove home in order to lift his morale. A 
standard patter developed between the young man and his father, 
who had himself been in business. 

Father: 'What have those boys in Brentwood tried on you this 
week?' 

Power: 'The guy is out for me. He wants my blood for 
some reason. He scares me. There's no other word I can use but 
fear.' 

Father: 'Stick it out. He's trying to test you. Stick it out. He 
must have a breaking point.' 

And Sugar did have a breaking point so far as Power was 
concerned. It came quite unexpectedly about six months after 
Power had joined Amstrad, in the run-up to the launch of the 
PC 1640 in June 1987. Power spent a weekend showing a prototype 
of the machine to Alan Fraser, a senior data processing manager 
for Ciba-Geigy, the Switzerland-based chemicals multinational, 
who was planning to review Amstrad's new computer in PC User. 
Power treated the Ciba-Geigy man like a king, giving him access 
to whatever software or technical expertise he needed as he tried 
out the PC 1640. Fraser's review, published soon after, was every- 
thing that Amstrad could have wished for: 'I would have no 
hesitation in recommending the machine in favour of the I.B.M. 
PS/2 Model 30 (I.B.M.'s latest flagship product) for corporate 
purchases,' it read. 

This ringing endorsement from a data processing professional 
in a large multinational fitted perfectly into Amstrad's continuing 



242 



ALAN sugar: the amstrad story 

campaign to break into the corporate market. When Power 
photocopied the review and showed it to Malcolm Miller, the 
marketing director took one look at it and told him to show it to 
Sugar. 

Sugar had by then bowed to the inevitable and moved into his 
own office, up a narrow stairway from the ninth floor. Power 
underlined the Ciba-Geigy quote, walked into Sugar's office, put 
the review on his desk, said, 'I think you'd better read this' and 
left. 

About 40 seconds later Sugar charged down the stairs, holding 
the review. He bounded up to Power and said, 'Have you done 
this? Is this your work? Are you responsible? This is bloody 
amazing!' Ecstatic, Sugar tried to tell the whole company about 
the review. He stood by the fax machine sending the review to 
Amstrad's agents in North America, Germany, Australia, every- 
where he could think of. 

From that day, Power was accepted; he would never receive so 
much praise from Amstrad's chairman again, but he was in the 
club. He went with Sugar to help with Amstrad product launches 
in the United States. Most important of all, Sugar began to treat 
him with respect. 

Power eventually left Amstrad in December 1987 to set up his 
own business. Like many others who have survived the Amstrad 
initiation, he emerged with a respect for Alan Sugar that verges 
on the embarrassing, 'I love the guy. I walked in there a smug 
young man of 22. And I had the crap knocked out of me for six 
months. I became a hard-nosed salesman and a professional. 
That was due to Alan Sugar. You go into the Alan Sugar school 
of management and you get the fat worked off you. He has this 
unbelievable ability to make marketing very simple,' says 
Power. 

In April 1987, at about the time when Power was winning his 
spurs, Alan Sugar set out his business philosophy in greater detail 
than ever before. By then Amstrad's chairman was being bom- 
barded with requests to explain in public the secrets of his success. 
He turned down almost all of them, concerned that speaking 



243 



DAVID THOMAS 

engagements would force him to take his eye off Amstrad's day-to- 
day affairs, but he made an exception when invited to lecture on 
entrepreneurship at the City University Business School - where 
he had held occasional discussions with the school's business 
students. 

Few lectures delivered in the world's business schools have ever 
been like that given by Alan Sugar. It had none of the glib 
formulae favoured by the management gurus who earn their crust 
advising other people on how to make a fortune. Amstrad's boss 
delivered a lecture peppered with jokes, anecdotes and Sugarisms, 
but it had one overwhelming advantage over the words of wisdom 
uttered by management pundits: the stamp of authenticity. Sugar 
had done it. He had made his millions by building up a hugely 
successful company and he had barely turned 40. 

The lecture hall was full of not only the business school's 
students and staff, but also analysts from London's nearby finan- 
cial centre. By all accounts, Sugar had his 300-strong audience 
eating out of his hands. Amstrad's chairman kicked off by giving 
budding multi-millionaires in his audience a few tips on what to 
avoid. Rule number one was not to pay too much attention to the 
City. Rule number two was to shun the press. 

'I could spend the whole of my working day talking to journal- 
ists, so much so that they have really become a pain in my life. I 
avoid them like the plague. The priority in life is to keep an eye on 
the business and not to get lured into the social high life, being 
exhibited around by the groupie type posers who wish to be seen 
with the new blue-eyed boy,' Sugar confided. 

With that out of his system, he set about the task of defining the 
essential components of Amstrad's success. 'Amstrad culture is all 
about realism, swift thinking and decision-making without com- 
mittees. Rise or fall by your own decisions or get out. A sense of 
urgency to get to the point.' 

Sugar tries to instill this sense of urgency into the people who 
work for him. 

In our company we attract people who either catch on very 
quickly or they last two minutes. When they catch on, they 



244 



ALAN sugar: the amstrad story 

understand the entrepreneurial flair of the company and see 
their colleagues using innovative ways and methods to achieve 
their tasks, not conforming to the standards that are written 
down in the books, but by cutting corners, taking a few risks, 
assuming the rest of the team will accept what they are up to 
when they have finalized their projects, not needing to stop step 
by step, in their thought pattern to see if the group agrees before 
they proceed. 

Amstrad culture is consciously aggressive. 'We have trained our 
staff to stand up and be counted. We don't want any corporate 
wimps. There are so many people in companies who shy away 
from confrontation, thinking that it will go away just like a bad 
cold; unfortunately that rarely happens.' 

The corporate bureaucrat has no place in Amstrad: There are 
those who are happy to be non-achievers in their life, who get 
buried in a big corporation and flit from job to job. They will 
never change their ways and, in a way, you have to admire how 
they get away with it all their life, without doing a day's work. 
There are others who are I suppose like me, who never expect 
anything for nothing and only know how to put one's head down 
and get on with the job. That is the type we at Amstrad attract.' 

Profits are the overwhelming focus of the Amstrad culture: 
'Amstrad is active in areas of profit only. We are not interested in 
producing a range of products just to put on a show.' Anything 
else would be vanity: There is no mileage to be known as the 
biggest producer in CTV (colour televisions). So what if 75 per 
cent of the range does not sell or yield profit?' 

Focus on profit underpins Amstrad's sales philosophy: 'When 
we have a hot selling item, you might think that we sit back on 
our laurels, admire our brainchild, watch it roll in the money for 
us, and sit around talking to one and all of the marketing success 
of the century. 

'Well, that's wrong! The Amstrad way is to thrash the thing to 
death. We look at the cost of the item and find ways to reduce this 
cost, so that we can make even more money while the product is 
selling. Our suppliers of parts and components of a hot item also 



245 



DAVID THOMAS 

think they are made for life and sit back thinking their boat has 
come in. We wake them up and alert them that all good things 
come to an end.' 

Amstrad has a distinctive attitude to successful and unsuccessful 
products: 'In most companies, it's normally with a product that 
doesn't sell that you sit down and find a way to make it cheaper. 
Not at Amstrad! A product that does not sell simply does not sell, 
so get out of it as soon as possible and wipe your face.' 

Looking coldly at a successful product is a key to making money: 
'Another Amstrad move is what the Japanese have termed v. a. It 
means value analysis. It is a nice way of saying knocking the price 
down. . . . 

'While all the blue-eyed boys in marketing are getting a pat on 
the back for the product, the v.a. boys are saying, for example, the 
power cord on that unit costs £1.50 at the moment, we use 50k per 
month, which is about 600k per year - which means we spend on 
that one item alone, £900k. Armed with that information, they 
study to see if they can reduce its specification and at the same 
time, look for other suppliers, and tell other suppliers that we can 
offer them, say, half the quantity and . . . what's the price?' 

Sugar explained that Amstrad might succeed in cutting the 
price of the power cord by 25p, yielding savings of £150,000 in a 
year. The best time to carry out this type of cost cutting is after it 
has launched a winner, because then it can hold out the prospect 
of large volumes when shopping around between alternative 
suppliers. 

At first, the reduced costs delivered by value analysis feed 
straight through into extra profits. But later, as a product matures, 
they can be used to cut the price of a product and thereby extend 
its life - while still generating the profit margins obtained when 
the product was first launched: Tor example, when a product was 
first thought of, we felt that at a retail price of £399 it will sell. If 
the product turns out to be a hot seller and you did the v.a. work, 
it is possible then to take the new v.a. cost and apply the original 
margin criteria and find that now a £299 retail price may be 
achievable. In doing so you may get some new life out of a product 



246 



ALAN sugar: the amstrad story 

which you think you have already thrashed to death. You can 
have a whole new re-launch campaign.' 

Sugar rammed home the point: 'The moral of this story is that 
if you have a winner, thrash it to death. Don't get complacent. 
Don't allow your head to swell. Knock the living daylights out of 
the thing, as opportunities come very seldom and you must make 
the most of things when the opportunity allows. I have seen it so 
many times in our industry that someone gets a hot item and they 
become untouchable. They feel their boat has come in and nothing 
can topple them from the cloud-cuckoo-land they are in.' 

Amstrad's approach to marketing is equally straightforward: 
Tor some reason, I have been called a barrow boy. I would take 
great exception to that if my ambition in life was to be seen daily 
at Annabel's with Lord and Lady Beseenwith. However, in a way 
marketing is just like a stall in Petticoat Lane. Frankly, it is no 
different. The owner of the stall is offering his or her wares. The 
sales pitch, albeit very rural and loud, is no different than some 
high cost advertising agency may apply. 

'The Amstrad approach to marketing is simple. We work on the 
philosophy - pile 'em high and sell 'em cheap. 

T suppose there was a big decision made in the early part of 
Amstrad life. The decision to be in the Bang and Olufsen (a 
Danish manufacturer of expensive audio and television products) 
end of the market or be in the Amstrad end of the market. 

'The B. & O. end of the market is not for me. Frankly, you 
spend all your life talking about the quality and performance, 
trying to hook the up-market yuppy by subtle remarks such as 
only the great and clever can understand. . . . Unfortunately this 
high quality hi-tech ramble has to be kept up all your life. You 
cannot slip back. You cannot be seen to enter the promotion end 
of the market, whereas it is easier going the other way.' 

Simple and direct advertising underpins this straightforward 
marketing philosophy. 'Philips will tell you they are simply years 
ahead. Ferguson will employ Andre Previn to tell you how 
wonderful the sound is. Toshiba use a misshapen Cockney robot. 
JVC have dramatic dancing girls. Amstrad show the product, 
mention the price and tell the punters where to buy it.' 



247 



DAVID THOMAS 



Amstrad's sole interest in launching an advertising campaign is 
to generate sales and thereby profits. 'It's incredible how other 
companies are far more interested in being seen to have the most 
catchy advert that everybody talks about, the advert that wins all 
the awards at the ad agencies' annual ball. However, they have to 
ask themselves - did it sell any product? 

T suppose it's an ego trip again. How does this happen? The 
company loses direction. The marketing manager is left to run riot 
on his own. The boss - i.e. the non-entrepreneurial boss - will 
assume his marketing man knows more than he does and allows 
him to get on with it. 

The Amstrad entrepreneurial method of marketing is a formula 
built up over the past few years, of trying to get value for your 
advertising money as soon as possible. We do not take the view 
that we should just spend a lump sum to promote the brand 
image. . . . We are happy to spend - and spend we do in big 
volume - but we want immediate return and at the same time get 
a branding rub off.' 

An Amstrad advertising campaign takes into account three 
factors: 'One - set the quantity of product, i.e. we plan to sell in 
the period xxx thousands of units, which will generate xxx 
thousands of gross margin. 

'The second factor is to alert the distribution and to sell in and 
make sure the goods are in the stores ready for the campaign. 

'The third factor is to produce an advert that gets to the bottom 
line immediately . . , 

'This philosophy is a great way to control the marketing people 
and lock them into results. It keeps your feet on the ground and 
brings home the reality you have to sell products that generate the 
profit in order to spend money advertising.' 

The Amstrad culture also colours its research and development 
activities: 'We have taught engineers to think on a commercial 
basis. This is quite a hard task. . . . 

Tn Amstrad, when the new custom chip that the engineer has 
designed comes off the production line, it will be used immediately 
in the mass production of the product for which it was designed. 

'Normally you wait and you test the thing - then you plan 



248 



ALAN sugar: the amstrad story 

production. As you can see, all this takes time and causes delay. 
The Amstrad way is to assume that the component is going to 
work and cut out the delay, but to have a contingency plan if it 
does not. 

'If some outside observers saw the risks that we take in 
engineering a new product, they would have kittens. But frankly 
there is no risk. Like all things in science it either works or it does 
not. If it works, we are months ahead. If it does not work, we are 
in the same boat as the conventionalists.' 

Summing up the Amstrad culture, Sugar took a tongue-in-cheek 
swipe at the pretentiousness and evasiveness of corporate slogans: 

Tan Am takes good care of you. 

'Marks and Spencer loves you. 

'I.B.M. says the customer is king. . . . 

'At Amstrad - We want your money!' 



Sugar's City University speech expounded the virtues of the 
entrepreneur as well as anything written in Britain during the 
1980s. A remarkable hymn of praise to the sharp, fast, aggressive 
values that distinguished the Thatcherite decade, it was also in 
many ways a quintessentially British performance. Business in 
Britain has always been marked by a lack of intellectual preten- 
sion. Pragmatic, common-sense virtues like those praised by Sugar 
in his speech appeal to British business people. British executives 
feel under less pressure than their counterparts in North America 
or Continental Europe to parade academic qualifications or an 
intellectual approach to their working lives. Business schools 
spread later in Britain than in most of its major competitors, 
notably the United States. Even by the late 1980s, qualifications 
like the M.B.A. (Master of Business Administration) were much 
less common in the top echelons of British companies than in 
corporate America. It is difficult to imagine even self-made 
entrepreneurs in North America or Continental Europe setting out 
their business philosophies in the raw style adopted by Sugar. 

An instructive contrast can be drawn with the account given by 
John Sculley of how he pulled Apple Computer back from the 



249 



DAVID THOMAS 

brink of financial disaster in the mid-1980s - in the process forcing 
out Steve Jobs, Apple's visionary founder. Sculley, who had been 
head of Pepsi-Cola before being lured to Apple by Jobs himself, is 
painfully concerned to convey his intellectual credentials in the 
600-page analysis of his experience at Apple (Odyssey: Pepsi to 
Apple, Fontana/Collins, 1987). 

When Jobs first visited his home, Sculley tells the reader, 
Apple's founder paused to admire Sculley's eclectic collection of 
books; 'not only on business and management but Zen, philos- 
ophy, architecture, art, and astronomy', Once Sculley had joined 
Jobs at Apple, the two men would spend a couple of hours each 
day swapping their thoughts. Jobs told Sculley that he would be a 
poet in Paris if he did not happen to work in computers. Entering 
into the spirit, Sculley confided in Jobs that he would undoubtedly 
have been an artist, if only he had not spent his life elbowing his 
way to the top of one of America's largest corporations. ; We 
tended to speak in half sentences and phrases, jumping from 
subject to subject: our pasts and our futures, marketing and 
technology, Trotsky and Marx, Coke and I.B.M.,' Sculley gushes. 

Much of this might be dismissed as an attempt by Sculley to 
convince Apple's long-haired computer wizards that he was not a 
grey man in a suit; but it is evident from his account that PepsiCo 
- his former home and the epitome of the methodical virtues of 
corporate America - was as pretentious in its way as Apple. 
Stonecutters were brought over from Europe to hand-cut the 
Italian cobblestones in the central courtyard of PepsiCo's 140-acre 
headquarters in New York state. The grounds are adorned with 
forty sculptures by some of the world's leading artists, while at the 
touch of a button on the desk of PepsiCo's chairman fountains 
shoot 40 feet into the air above a P-shaped, man-made lake. 

It is no wonder that the reporter from Time .was taken aback by 
the stuffing coming out of Amstrad's chairs. In the cultural 
vacuum of the United States, business can invest itself with all 
manner of intellectual and cultural pretension. In Britain, by 
contrast, business has not developed an intellectual self-image. 
While it is debatable whether this has been to the disadvantage of 
British business, Sugar for one has never made any secret of his 



250 



ALAN sugar: the amstrad story 

contempt for the flabbiness he detected at the heart of many large 
American corporations. 

'I have had a bellyfull of Americans who flit from job to job and 
exist in large empires as statistics - the sales people living in a 
dream world, asking you to build 50 factories on the basis of their 
potential customers saying, "Have a nice day." They move around 
from job to job, usually getting found out for what they are,' he 
told his City University audience. 

Ironically, Sugar delivered his City University speech just as 
Amstrad was in the process of abandoning crucial ingredients of 
its ultra-entrepreneurial, small business ethos. Few people realized 
it at the time, but in many ways the address set the seal on the 
first twenty years of Amstrad's existence. Indeed, Sugar had 
already sown some of the seeds which would grow into structures 
more appropriate for a large company. 

The speech contained a clear recognition of the dilemma facing 
Amstrad: 

There is a danger that exists in my company that as we grow, 
we need to put in place more layers of management, control 
systems, additional departments and a lot of other paraphernalia 
which are considered necessary in a big corporation. The 
problem with this is that the whole machine can lose its speed 
and direction. . . . All companies, especially in the hi-tech field, 
must strive to grow - but growing too big too fast makes a 
bureaucracy and risks the flair and culture stagnating. 

In May 1985, Sugar had brought Ken Ashcroft into Amstrad as 
its new finance director. Then just entering his fifties, Ashcroft 
came with plenty of experience in the finance functions of large 
companies, thanks to a career which had included stints at Ford, 
Philips, Comet, Hepworth and Dixons. He had first caught Sugar's 
attention through his work at Comet in the early 1970s, when he 
introduced a computerized planning system for the buyers which 
gave them precise information about the state of their inventory, 
orders and cash availability. The quality of information this system 



251 



DAVID THOMAS 



generated allowed Comet to be much less volatile than other 
retailers in its dealings with suppliers like Amstrad. 

Ashcroft's speciality was helping companies which had out- 
grown their existing financial control systems, and that was the 
job which Sugar wanted him to perform at Amstrad. When he 
joined, Amstrad's senior managers were not receiving the operat- 
ing information which is standard in large companies - monthly 
data about sales, inventories, margins and so on. Instead, Sugar 
still had sufficiently close oversight of all the company's business 
to be able to produce an accurate stab at a six-month forecast after 
spending half an hour with a calculator. However, this situation 
could not continue much longer. Ashcroft's first priority was to 
introduce a computer-based reporting system which locked 
together financial information with data about the physical loca- 
tion of Amstrad's products; this gave the senior managers instant 
access to information on the company's current financial position, 
as well as on its immediate prospects. 

Introducing these systems required Ashcroft to build up finance 
and data processing departments at Amstrad, though Sugar's 
natural instincts were against such developments. 'Why are you 
spending my bloody money on overheads?' would be Sugar's first 
reaction. He argued against a proposal and called it stupid, yet he 
also listened. And if he was convinced, he would agree. 

Ashcroft's prodding also prompted Sugar to introduce a pro- 
fessional treasurer to manage the flow of cash through the com- 
pany. Before Ashcroft arrived, Sugar had been doing this himself. 
A key part of treasury management in Amstrad, as in most 
companies with business in more than one country, was to deal 
with foreign exchange matters. It bought many of its supplies in 
the Far East in yen or dollars, sold its finished products in Europe 
and then reported its results in sterling. The Amstrad chairman, 
with his unrivalled experience of trading in the Far East, proved 
adept at knowing when it was sensible to borrow in yen and when 
in sterling. The currency function was never a formal profit centre, 
as it is in some large companies, but Amstrad did fairly well from 
it. 

Sugar was readily persuaded to relinquish day-to-day currency 



252 



ALAN sugar: the amstrad story 

management to a full-time professional treasurer in 1987, yet he 
continued to suggest innovative approaches in this area. For 
example, in the late 1980s he made the unorthodox suggestion 
that Amstrad should carry out much of its borrowings in yen. 
Borrowing in strong currencies like the yen carries a substantial 
risk: as the yen appreciates, the sterling value of the loan mounts 
correspondingly. But this strategy also has benefits: Amstrad could 
take advantage of the lower interest rates attached to yen borrow- 
ings and Sugar calculated that this interest rate differential was 
more than enough to offset the risk of currency appreciation. 

It was a strategy which gave an orthodox finance director like 
Ashcroft a few sleepless nights, as he recalls: 

I must say that borrowing in a strong currency, taking advan- 
tage of lower rates, isn't something that appeals to me as a 
finance guy. It's very risky. But Alan pointed out that there was 
such a huge gap in the forward rates that the yen would have 
had to strengthen a heck of a lot before we needed to take action. 
And a jolly good strategy it's been, because the yen has oscillated 
quite violently, so we've found we've been able to buy our 
requirements in yen when it's been relatively weak against the 
pound. 

Ken Ashcroft joined a team of board members who were Sugar's 
immediate lieutenants. Malcolm Miller was in charge of sales and 
marketing. Bob Watkins oversaw the development of new Amstrad 
products. Jim Rice was in charge of operations such as manufac- 
turing and warehousing. It was a tight team, carrying no passen- 
gers and with Sugar the clear leader, as he explained in early 1986: 
'Businesses need leaders, spearheads. Some people pad out their 
boards with right honourables and air marshals and fancy titles 
like that. But they still need spearheads if they want to succeed. I 
provide Amstrad with its prime motivation.' 

Yet as Amstrad continued to grow in the second half of the 
1980s, inevitably it had to take on more and more staff. By the end 
of the decade, there were about a dozen people working on sales 
and marketing at Amstrad headquarters under Malcolm Miller. 



253 



DAVID THOMAS 

Bob Watkins and Richard Altwasser, the former Sinclair computer 
designer, gradually built the firm's in-house research capability 
into a team of fifteen engineers. Ken Ashcroft's finance and data 
processing teams grew to about fifty, although many of these were 
engaged in routine clerical functions. 

Even as it grew, Amstrad managed to retain a lack of formality. 
Ken Ashcroft always impressed on the team of auditors from 
Touche Ross during their annual visit that they should come 
straight into his office if ever they had any problems. The 
accountants were used to more bureaucratic arrangements with 
their clients: typically, problems would be referred up the line 
through a client's finance department until eventually they found 
their way on to the finance director's desk. But at Amstrad Ken 
Ashcroft's door was always open. Indeed Eric Tracey, the Touche 
Ross partner in charge of the Amstrad audit, once told Ashcroft 
half-jokingly that he ought to try closing his door now and again. 
'With my chairman, if I closed my door, he would tear it off its 
hinges,' was Ashcroft's instantaneous reply. 

Amstrad was still a lean company if staff numbers were com- 
pared with the size of its turnover, yet it was no longer the skeletal 
organization it had been in the mid-1980s. Malcolm Miller 
describes the impact in sales and marketing: Two or three years 
ago, it was just me. Me and maybe one other. Over the past three 
years it's had to change, because the products have become so 
varied and complex, and the budgets have become so much 
bigger.' 

The inevitable result of this rapid expansion was that Amstrad 
filled up with new young faces. The company had a clear idea of 
the type of newcomer it was seeking; Malcolm Miller, who 
emerged in the 1980s as the most visible person in Amstrad after 
Sugar, explains: 'We've taken young, bright graduates. But more 
importantly they have to have the right culture. They have to be 
entrepreneurial and enterprising. They have to be able to rewrite 
the rule book, if necessary.' 

As Miller's key sales and marketing function expanded, he 
introduced more structure into the operation, focusing his team 
on individual products and brands along the lines he had learnt 



254- 



ALAN sugar: the amstrad story 

as a young trainee in the Birds Eye division of Unilever: 'Really, 
I've taken a lot of it from the brand manager position at 
Unilever, whereby the person is the product. He eats, drinks, 
sleeps the product from the day it is born to the day it dies. He 
analyzes the sales trends, the seasonality, the packaging, the 
pricing, the competition. He's got to know everything. He is the 
product.' 

But Miller was careful not to let bureaucracy slip in along with 
the greater structure: 

If you haven't got a sense of urgency and entrepreneurial flair 
and you just want to be a bureaucrat pushing people and paper 
around, it doesn't work here. We will not become like the people 
who sit on committees all day long, pushing papers around, 
theorizing about some activity which might increase the brand 
share by 0.1 per cent to raise at next month's marketing 
committee. That's not us. 

The formal meetings held within the sales and marketing depart- 
ment are tightly circumscribed, as he says: 

Every Monday, we have a sales and marketing management 
meeting, where we talk about sales on a monthly basis and 
discuss the problems and opportunities on a weekly basis. 
Monday for an hour. We have these rules that a meeting 
shouldn't last longer than an hour. We have another rule that if 
there's somebody important on the phone, you take that call. 
You don't say I'm in a meeting, because the call might be more 
urgent than the meeting you're in. 

The influx of newcomers inevitably meant that Amstrad began to 
employ people with a relatively shallow commitment to the 
company. Miller describes how new members of staff are inducted 
into the Amstrad culture: 

We allow them to make important decisions, which they 
couldn't do in a larger company. We allow them to feel part of 



255- 



DAVID THOMAS 

the problems and opportunities, successes and failures we have 
every day. Certain people like responsibility. They like to feel 
they can change the face of one of Amstrad's projects - they can 
do a deal with a big software company on a promotion; they can 
come up with a new creative idea for an advert; they can find a 
new retailing opportunity that makes a big difference. 

Dealing direct with Amstrad's boss is also an experience that can 
make or break the young bloods, as Miller explains: 

Now and again, I let Alan loose on them. That experience is 
always good. They always find it very stimulating. I say to 
them, 'Let's show Alan this idea of yours.' He'll sit and 
deliberate and they'll get a chance to argue. Some of them will 
argue, but if we don't like the idea we say so immediately. 
Nothing gets hidden here. People are very straight talkers. If 
you don't like straight talking or you're shocked by being told 
your idea is crap, then don't come to Amstrad, because we don't 
pull our punches. 

As Amstrad prospered, Alan Sugar hit on a more direct means to 
encourage the dynamism and loyalty of his staff. Share options are 
a way of rewarding a company's work-force for the contribution 
they have made to its success. Typically, a share option scheme 
introduced on a particular date - call it x - allows the staff to buy 
shares at a future date, for example x + three years, at the share 
price which existed at x. If the company has prospered and the 
share price has risen during the three years, then those covered by 
the scheme can make substantial capital gains on realizing their 
options. 

Amstrad introduced its first share option scheme on 23 July 
1985; it allowed those in the scheme to exercise their options three 
years later at a price of 13.6p a share (the mid market price on 23 
July 1985, after allowing for the subsequent fivefold split of the 
shares in June 1986). This is a standard format for such schemes. 
The relatively unusual feature of Amstrad's scheme was the 



•256 



ALAN sugar: the amstrad story 

generous treatment it gave to a large number of the company's 
staff. 

In total, 51 directors and employees of Amstrad were included 
in the company's first share option scheme, about a tenth of the 
work-force at the time of introduction in 1985. Alan Sugar ensured 
that all those who had helped him build up Amstrad were in the 
list, including five Chinese members of the company's Hong Kong 
staff. Options to buy a total of 11.551 million shares were 
distributed to these 51 people. On 25 July 1988, the first day on 
which the options could be realized (allowing for a weekend), 
Amstrad's share price stood at 215p. The 51 beneficiaries were 
sitting on a potential capital gain of £23.26 million (11.551 million 
shares X 215p-13.6p) - almost £500,000 each on average. 

This gives only a very general indication of how each of the 5 1 
individuals was affected by the scheme. For a start, the 11.551 
million shares were not distributed equally; the options ranged 
from the 1 . 1 76 million allocated to each of the five Amstrad board 
members in the scheme (Ken Ashcroft, Malcolm Miller, Jim Rice, 
Marion Vannier and Bob Watkins), to the 73,500 shares allocated 
to the least senior staff in the scheme. In addition, those covered 
by the scheme could choose to wait until after 25 July 1988 to sell 
their shares, which would have brought them either higher or 
lower gains depending on the share price at the time. In fact, the 
five board members all made sizeable disposals as soon as they 
could; they sold between 115,000 and 400,000 shares each on 25 
July 1988, yielding capital gains ranging from £230,000 to 
£805,000. 

The five Amstrad board members and two other employees 
stood to gain over £1 million each as a result of this first Amstrad 
share option scheme. In other words, seven new Amstrad million- 
aires had joined Alan Sugar. The most junior staff stood to gain 
almost £150,000 each. Amstrad also distributed smaller numbers 
of share options to staff in later years. Alan Sugar was no longer 
alone in having benefited handsomely from the company's 
success. 

At Amstrad's annual general meeting in 1988, institutional 
investors questioned Sugar about the share options, which they 



257 



DAVID THOMAS 

saw as unnecessarily generous. The chairman would have no truck 
with these criticisms and defended the policy vigorously, arguing 
that total dedication by the staff was integral to Amstrad's success 
and needed to be adequately rewarded. 

Amstrad did without much of the rest of the paraphernalia of 
professional staff relations to be found in more typical companies 
of its size. Unions, for example, were an anathema, as indeed was 
the whole bureaucracy of personnel management. One head of an 
Amstrad subsidiary even had to pretend that its personnel officer 
did an entirely different job, so fearful was he of the scorn that 
Sugar would pour on him for carrying such a (in Sugar's eyes) 
useless post. Sugar also thought that Amstrad could do without a 
gold-plated pension scheme of the kind typical in most large 
businesses. 

While expecting hard work from its employees, Amstrad does 
not look for empty demonstrations of devotion. This shows through 
in its attitude to working hours. Unlike some other high-achieving 
organizations, no Brownie points are earned at Amstrad simply by 
burning the midnight oil beyond the official end of the working 
day. 'Even today, if you stand in front of the exit at a minute to 
five, you get trodden into the carpet,' one Amstrad insider says. 

This attitude flows partly from the top, since Sugar has always 
made a point of trying to get home to his family at a reasonable 
hour. His instincts have changed little since 1980 when, at the 
time of Amstrad's flotation, he told a reporter, 'I take my holidays 
and I think five days' work is enough. If a job isn't finished on 
Friday, it can wait until Monday.' 

Running the large company which Amstrad has become inevi- 
tably increases the demands on Sugar, but efficient time manage- 
ment is his way of cramming an abnormal work-load into a normal 
working day. Most days he just sits in his office working through 
whatever is at the top of Amstrad's agenda. He usually skips lunch 
or, at most, eats a sandwich at his desk. During meetings with 
staff, he often devotes half his brain to clearing through his 
paperwork or to completing a few deals on the phone. Most 
meetings with outsiders are tightly structured, as Paul Bailey - a 
software supplier to Amstrad - describes: 

•258- 



ALAN sugar: the amstrad story 

All Sugar's meetings are very much to the point. They have a 
fixed agenda and just enough time budgeted to the exercise to 
get to a closure point. He very rarely passes the time of day in 
his meetings. Within the first few minutes the objective is clearly 
established, so each side knows what it is bargaining for. Sugar 
tends to lay down his terms quite early in a meeting, so it will 
be up to the other side to say whether they are willing to agree. 
The other side can move the terms, but usually not by a whole 
lot. He has an incredibly focused mind. 

Perform well or clear out: that is the hard injunction at the centre 
of Amstrad culture. But its contempt for corporate wimps, bureau- 
crats and timewasters is not simply macho breast-beating. 
Amstrad culture has a laser-like concentration on sales and profits: 
it is focused, like Sugar himself. 



259- 



• 14- 
Birth of a Multinational 



In 1987 Amstrad set out on a new path which was to propel it far 
from its small business roots. Breaking with his previous practice, 
Sugar created a string of overseas subsidiaries - a policy departure 
which amounted to the re-fashioning of Amstrad into a true 
multinational corporation. 

The impetus for this new strategy came from the policy review 
which Amstrad had carried out in late 1986 and early 1987, partly 
with the help of the Boston Consulting Group. Sugar concluded 
that Amstrad did not need to march into radically new businesses 
such as white goods. There was still plenty of progress to be made 
by building on its existing product lines and by extending the 
geographical spread of its activities, even if growth was bound to 
be slower than in the heady years of the mid-1980s. 

Alan Sugar gave a characteristic account of the new strategy in 
his chairman's statement accompanying the 1986-87 accounts: 

There is a temptation to follow the experience of other com- 
panies which have risen, like Amstrad, to great heights. The 
phenomenon of meteoric growth followed by the question, 'What 
do we do for an encore?' The answer is often that the company 
wanders like a lost lamb with a shopping basket, surveying the 
shelves of the company lists in desperation, buying profits at a 
high cost to cloud the ongoing issue. 



260 



ALAN sugar: the amstrad story 

That would not be Amstrad's way. Instead, Sugar had decided, 
his company would build up its presence abroad. This required a 
fundamental re-orientation of its overseas activities, as he 
explained to the shareholders: 

Amstrad will no longer distribute its products in major overseas 
markets through agents who, in many cases, also represent other 
manufacturers. Such arrangements lead to the dilution of the 
sales and marketing effort. In addition there is no longer any 
room for a 'middle man's' margin. The group can now offer the 
resources to overseas subsidiaries, enabling them to penetrate 
their markets in the way we have done in the U.K. We must 
control our own destiny and to do so we must control our 
marketing and advertising effort in each market, as well as 
product planning. 

Mixed experiences overseas lay behind these statements. Before 
1987, Amstrad had carried out all its activities abroad through 
agents except in France - an anomaly explained by the special 
circumstances of Marion Vannier, boss of Amstrad France — and 
in Hong Kong, which was a co-ordinating point for the company's 
interests in the Far East. Yet only in France and Spain had the 
sale of Amstrad's products begun to match its success in Britain. 
Every country presented its own problems, but in each case there 
seemed to be a common solution. 

The United States, for instance, is a market which Sugar has 
consistently approached with great caution. Notoriously, America 
has been a graveyard for British computer companies. In the first 
half of the 1980s Acorn, Apricot and Sinclair Research launched 
into the U.S. with great fanfare. They made initial gains and were 
then seriously burnt when they found their cheque books too small 
to take on the giant American personal computer market, particu- 
larly when it went through one of its periodic recessions. Learning 
from their mistakes, Sugar looked for someone else to bear the 
costs of marketing and distributing Amstrad computers in the 
U.S., just as he had done in Spain. His aim was to leave Amstrad 
unexposed if its products did not sell well in North America. 



261 



DAVID THOMAS 



'Everyone in America seems to think that the only way to break 
into that market is to spend $200 million on a big national ad 
campaign. That doesn't fit with our first priority - not losing any 
money there,' Sugar explained. 

Amstrad achieved what appeared to be a breakthrough in 
March 1986, when it signed a marketing agreement with Sears 
World Trade, the overseas trading arm of Chicago-based Sears, 
Roebuck - the world's largest retail chain. Sears was bullish at 
first, telling the Wall Street Journal that it expected Amstrad to sell 
more word processors in North America than in Europe. It seemed 
set to place an initial order for 100,000 of Amstrad's PCW8256 
word processors for sale in the U.S. and it looked as though Sugar 
had hit the jackpot: a costless route into the North American 
market on the back of one of the country's most powerful retailers. 

In the event, relations between Amstrad and Sears quickly 
soured. By October 1986, Sears had taken delivery of only 70,000 
of the 100,000 word processors which Amstrad was expecting to 
sell it. Worse, the giant retailer appeared to be doing nothing 
much with the machines it had. Sears was not selling Amstrad's 
products in any quantity. Part of the trouble was that the 
American public insisted on viewing the PCW8256 as a not very 
powerful CP/M-based computer, rather than as a dedicated word 
processor. The standards-conscious U.S. market was also discon- 
certed by Amstrad's non-standard 3" disk. 

While it would probably always have been an uphill struggle to 
sell many Amstrad word processors in North America, Sugar also 
became convinced that much of the problem lay with Sears. He 
aired his feelings six months later in his City University speech: 

In the U.S.A., we sold goods to Sears - the mighty Sears, where 
the left hand does not know what the right hand is doing. They 
obtained the marketing rights to our word processor, ordered 
shiploads and stored them in a warehouse in stacks nearly as 
high as the Sears Tower in Chicago. Someone forgot to organize 
the marketing. 

The machine the Sunday Times called the marketing miracle of 
the decade stayed in stock as the peak selling season came and 



262 



ALAN sugar: the amstrad story 

went. We had been paid for the product but we missed the 
opportunity of exposing our brand. In short, we went to the 
wrong people. We went with the buying arm of the world's 
largest retailer. They were great buyers but they could not sell 
anything. This is a case of big not being beautiful. The Sears 
Tower has 100-odd floors and I think they have that many 
layers of management. 

A few months after the word processor deal, Sugar flew on 
Concorde for a meeting in New York's John F. Kennedy airport 
to consider U.S. marketing plans for Amstrad's I. B.M. -compatible 
PC 1512 personal computer. Already he was thinking of ditching 
Sears. Three or four executives came to the meeting from Sears's 
headquarters in Chicago. A couple of people from the U.S. 
software company, Digital Research, which supplied much of the 
initial software for the PC 1512, were also there, including its 
president. Sugar softened up the Sears team by boasting that he 
had set out for the meeting from Brentwood after they had left 
Chicago. He then took them through the PC 15 12 and his sales 
hopes for the computer in the U.S. 

The Sears executives raised doubts about the PC1512's chances 
in North America: they could not see how a new entry like the 
PC 1512 could hope to make a big inroad into the U.S. which, 
unlike Europe, had plenty of low-cost I. B.M. -clones in the market- 
place. As hesitation was heaped upon hesitation, those present 
could see Sugar grow tense, visibly frustrated with the cautiousness 
of the Sears response. 

After the meeting was over, Sugar drew aside the two men from 
Digital Research and told them not to worry about the impasse 
with Sears; he was working on alternative distribution plans for 
the U.S. Sears's attitude at the J.F.K. meeting had been the last 
straw, and shortly afterwards Amstrad signed up a new distribu- 
tor, Video, which until then had been a leading U.S. distributor of 
Commodore computers. Vernon Moore, Video's president, sev- 
ered his links with Commodore and set up as Amstrad's U.S. 
agent in Dallas, Texas. Amstrad entered 1987 with Video handling 
its North American business, but still with no financial exposure 



263 



DAVID THOMAS 

in this tricky market. Video paid for all the machines it ordered in 
advance, and bore all the marketing expenses. 

However, it soon became plain that Video did not have the 
financial strength to make much of a dent on the North American 
market. Amstrad would have to take the reins if it wanted more 
than negligible U.S. business. In September 1987, Amstrad estab- 
lished its own subsidiary in the U.S. by acquiring Video for $7.5 
million - $2 million in cash and 2.2 million Amstrad shares. 

While Amstrad struggled with Sears during 1986, over in 
Madrid Jose Luis Dominguez had been considering his next move. 
He was wondering how to cash in on his tremendous success in 
launching the Amstrad brand in Spain. The obvious route was to 
float his distribution company, Indescomp, on the Spanish stock 
exchange. Late in 1986, Dominguez phoned Sugar and told him 
about his plan. The Amstrad chairman was non-committal, so 
Dominguez began to take the first steps, but in early 1987 Sugar 
called him over to Amstrad\s headquarters. He explained to the 
Spaniard his plans to strengthen Amstrad's control over its 
overseas operations, and said that an independent Indescomp 
would not fit this new strategy. Rather than floating Indescomp in 
Spain, why not sell it to Amstrad? 

Dominguez hesitated about losing his independence, but in the 
end he conceded, negotiating what turned out to be a more than 
good price. The deal was announced in September. Amstrad 
bought Indescomp, renamed it Amstrad Espafia and appointed 
Dominguez head of its new Spanish subsidiary. It paid £21.65 
million for Indescomp - made up of 8.812 million Amstrad shares, 
worth about £17 million, with the rest in cash. Another Amstrad 
multi-millionaire had been created. 

Sugar also had to develop operations in West Germany, a 
country where Amstrad had been selling some goods since the 
early 1980s. An electronics company called Schneider Rundfunk- 
werke, run by the two Schneider brothers, was the driving force 
behind the company's entry into West Germany. After noticing 
Amstrad's equipment at an audio fair in London in 1980, Bern- 
hard Schneider approached Sugar and offered to sell its equip- 
ment in West Germany. But at the time Sugar was reluctant to 



•264 



ALAN sugar: the amstrad story 

enmesh Amstrad in the expense of tackling the West German 
market. A deal was therefore agreed by which Schneider bore all 
the costs and risks of selling Amstrad goods in West Germany. In 
return, Schneider's name appeared on the products, a decision 
which Amstrad was to regret. 

The first Amstrad product to sell strongly in West Germany 
was the CPC464 home computer; the West German market 
contributed £32.5 million to Amstrad's sales in 1984-85, almost a 
quarter of the company's total turnover. So well regarded was the 
CPC464 in West Germany that it was paid the compliment of the 
sincerest form of flattery - imitation. A German outfit thought the 
ROMs controlling the CPC464's add-on disk drive sufficiently 
elegant to copy. Roland Perry, Amstrad's technical manager, flew 
to Munich to testify in the resulting copyright case - the only 
known example of someone trying to clone an Amstrad design. 
Amstrad won the case. 

But relations with Schneider began to cool when the German 
company failed to make a success of the PCW8256 word processor. 
In Sugar's eyes, Schneider committed two cardinal errors. First, it 
promoted the PCW8256 as a computer rather than as a dedicated 
word processor. Second, Schneider marked up the price of the 
PCW8256, pitching the price at what it thought the market would 
bear and ignoring the Amstrad maxim of creating a mass market 
by setting a low price based on the costs of production. 

Alan Sugar explains the point: 

The Germans took the Joyce (PCW8256) too seriously as a 
computer. That's why it was a failure in Germany. They 
marketed it as a computer - and put too high a price on it. If 
they had come into the market with that product at the 
equivalent of £399 in D-marks and presented it as a word 
processor, a replacement for a typewriter, they would have had 
a flaming success. But they entered the market at nearly double 
the price, thinking they could soak up a lot of profit. But of 
course the German market said, 'Why should we buy this for 
£700 when we need spend only another couple of hundred 
pounds to get an I. B.M. -compatible computer?' It was a wrong 



265 



DAVID THOMAS 

attitude to pitch the price near the competition. What we needed 
in Germany was a blockbuster, something to make people look 
up and say, 'My God, that's cheap!' 

Bernhard Schneider, in charge of Schneider's sales and marketing, 
disputes this account: 'I'm not really sure why the PCW did not 
sell so well in Germany,' he says. 'Maybe we in Germany are a 
little bit more conservative about such things than people in 
Britain. Perhaps even now the Germans prefer to use a typewriter 
to a computer. Britain is two years ahead of Germany in that 
respect.' 

Whatever the reason, Amstrad's sales in West Germany stag- 
nated in 1985-86 at £37.5 million, barely up on the previous year. 
Greater success with the I. B.M. -compatible PC 1512 helped 
Amstrad's German sales to jump to £63.9 million in 1986-87, a 70 
per cent increase. With its sales flourishing in other Continental 
countries, Sugar had become determined to build up Amstrad 
brand awareness in West Germany: under pressure, Schneider 
had agreed to put 'Schneider by Amstrad' on its PC 15 12s. By the 
time Sugar wrote his chairman's statement in September 1987, 
however, the days of dealing with Schneider were numbered. 

The German company was becoming keen to branch out on its 
own. Based near the spa town of Bad Woerishofen in rural 
Bavaria, Schneider is blessed with a touch of Amstrad entrepre- 
neurialism. The Schneider brothers, increasingly uneasy about 
acting as middle men for Amstrad, wanted to manufacture their 
own computers. 'Under our old situation, Alan Sugar controlled 
everything and we had to follow him,' Bernhard Schneider says. 

Schneider was chafing under the terms of its deal with Amstrad, 
which restricted its computer operations to the German-speaking 
countries, Belgium and the Netherlands. The German company 
had learned from Amstrad how successful a fleet-footed consumer 
electronics company could be if it controlled its own design, 
sourcing and marketing. By the end of the 1980s, after severing 
relations with Amstrad, the Schneider brothers had built up their 
firm to annual sales of about DM1 billion (£350 million). 

From Amstrad's viewpoint, the break with Schneider came at 



•266 



ALAN sugar: the amstrad story 

an opportune moment, since it was set on the path of creating its 
own subsidiaries. Head-hunters scoured Germany in 1987 looking 
for someone to run Amstrad's subsidiary there; Sugar had told 
them to pick an executive who was number two in an existing 
computer company and therefore hungry to succeed. The head- 
hunters recommended Helmut Jost, then Commodore's second-in- 
command in Germany. He was given the job of setting up from 
scratch Amstrad's German operation, which became fully oper- 
ational in 1988. 

During 1987, Amstrad formed subsidiaries in two countries 
where its presence was not as yet strong - Italy and Australia. 
The ending of the Schneider deal also prompted Amstrad to create 
companies in Belgium and the Netherlands, where Schneider 
previously handled sales of its equipment. Within the space of a 
year, Sugar had formed subsidiaries in seven countries: Australia, 
Belgium, Italy, the Netherlands, Spain, the United States and 
West Germany. Amstrad would still make opportunistic sales 
outside these territories. At one point in 1987, for example, 
Amstrad's computers accounted for 70 per cent of the air freight 
between London and Warsaw, as it responded to a surge in Polish 
demand for personal computers. But Amstrad now had its own 
operations in place in most of the world's main markets. 

While Sugar was busy setting up this marketing network 
overseas, Amstrad's Hong Kong operation had been going from 
strength to strength. The numbers employed there by Amstrad 
grew steadily to a peak of 428 in October 1988. A total of 370 
worked in its factory, which was sandwiched into four floors of a 
large commercial building in Kowloon. The remainder were 
engineers or administrators working in Amstrad's offices, tucked 
away in a bustling jewellery quarter. Amstrad still shared its 
offices with the same textile company which had offered Randall 
the use of a small room when he opened the Hong Kong operation 
back in 1981. By the end of the 1980s, however, Amstrad dwarfed 
the textile company, whose employees could be seen poring over 
materials in a corner of Amstrad's offices. 

By then Amstrad Hong Kong had created a sophisticated 
system for overseeing the company's sub-contractors dotted 



267 



DAVID THOMAS 

around the Far East. Typically, up to five of its Hong Kong 
engineers would work full-time in a sub-contractor's factory when 
a new product was being introduced. Amstrad supervised and 
paid for the installation of new tooling for its products. Once a 
sub-contractor was bedded down, it maintained pressure on 
quality through a team of twenty inspectors - based in Hong Kong 
- who visited its sub-contractors continuously. 

Amstrad was loyal to its long-term sub-contractors, like Orion, 
but it was part of Randall's job in Hong Kong to look out 
continuously for cheaper sources of supplies - to 'cherry-pick', as 
Alan Sugar described it, from the best available in the Far East. 
'You always have to keep your options open to move your 
production round the region,' argues Randall. 

Amstrad needed to keep an open mind about its suppliers 
because of the changing fortunes of the different Far East econ- 
omies. Gradually it reduced the quantities of products and com- 
ponents it sourced from Taiwan as labour costs rose there. It kept 
a close eye on the rising currency, political instability and labour 
cost pressures in South Korea, where most of its computers were 
made. Like many other electronics concerns in the Far East, 
Amstrad also began to source some goods from China; it had 
components such as its computer mouse made in factories in 
Guangdong, the Chinese province next to Hong Kong, tapping 
China's abundant supplies of cheap, flexible labour. 

On one occasion, Randall was impressed by the goods coming 
out of a Taiwanese factory which Amstrad had not used before. 
Its Hong Kong engineers took the factory's goods apart, studying 
the techniques used by their Taiwanese counterparts. Then they 
worked out a price and quality specification for a large quantity of 
a particular Amstrad product. 'At that stage, the factory didn't 
even know we w r ere interested in them,' Randall says. 

When Amstrad knew what it wanted and the price it would pay, 
its team visited the Taiwanese factory and spent four days crawling 
over its procedures. Satisfied with what they found, they began 
negotiating an order for tens of thousands of units. 

Every month Amstrad's Hong Kong office received bills of 
materials from its sub-contractors. Its staff examined these, trying 

•268- 



ALAN sugar: the amstrad story 

to spot components which Amstrad, with its large purchasing 
muscle, could buy more cheaply. This effort was helped by a 
central purchasing department at the British headquarters which 
also monitored prices worldwide. 

Amstrad's Hong Kong factory was not just the place where 
Amstrad made printers, which sold in similar numbers to its 
computers. It was also used as a testing ground for new production 
ideas. If these ideas worked, then Amstrad would encourage its 
sub-contractors to follow suit. 

Randall's ability to motivate his young Hong Kong Chinese 
staff was crucial to his operation, not least because engineers and 
technicians there are in chronically short supply. He gave his staff 
responsibility and recognition as early as possible. The best 
example of this is Callen So, who was promoted from being a 
secretary to become Amstrad's sales and marketing director in the 
Far East and then brought to Amstrad's headquarters as special 
assistant to Alan Sugar, with a desk next to his. Another case was 
Isaac Ip, a young Chinese man who joined the company as a 
junior draughtsman. Randall encouraged Ip's early interest in 
design, with the result that Ip came up with a more efficient mouse 
for Amstrad's personal computer line. Ip, in turn, was promoted 
to the number two position in the engineering department in Hong 
Kong. 

Married to a Chinese woman, Randall had little time for the 
rituals of the expatriate British community; he divided his energies 
between workaholic devotion to Amstrad and his small sailing 
boat, moored in the harbour of the fishing village where he lived. 
He believed that Amstrad's trust in its key Hong Kong staff paid 
dividends. It not only helped to pick up commercial gossip in the 
surprisingly small world of Far Eastern electronics sub-contrac- 
tors, but also came in useful when Randall prepared for nego- 
tiations with Japanese suppliers: he went in for role-playing 
exercises, with his Chinese lieutenants pretending to be his Japan- 
ese adversaries. 

Alan Sugar typically came to the Far East, usually with 
technical director Bob Watkins, to handle the main annual 



269 



DAVID THOMAS 



negotiations with the company's biggest sub-contractors. Bargain- 
ing took place in intensive sessions, unsuitable for anyone weak in 
mind or body. Once when Callen So went with Sugar and Watkins 
to see Funai, one of Amstrad's main Japanese suppliers, she had 
hoped to spend an hour or two looking at the Japanese shops. But 
no such luck. The Amstrad team was picked up by a Funai car 
from the hotel lobby each day at 7.30 am and taken to the factory, 
where they held face-to-face negotiations with about a dozen 
Japanese managers and engineers right through until 6 pm, with- 
out a break for lunch. The two sides argued over the benefits of 
using a particular component; about whether Funai would shave 
a few more cents off the price if Amstrad ordered large volumes; 
and, crucially, over the performance level that could be delivered 
for a given price. 

Yet in the second half of the 1980s, a range of pressures emerged 
which was to endanger the stability of Amstrad's elaborate system 
of sub-contractors in the Far East. Labour costs and exchange 
rates began to rise, notably in Taiwan and South Korea (they had 
already done so in Japan), throwing into doubt the core cost- 
cutting reasons for Far East sourcing. A growth of protectionist 
sentiment within the European Community also threatened 
Amstrad's network of relationships in the Far East. 

This sentiment was channelled into 'anti-dumping' cases by the 
European Commission against Far Eastern producers. Under 
international trading rules, anti-dumping duties can be levied 
against an imported product if it is sold for less than in its home 
market, or at less than the cost of production plus selling costs and 
profit. From 1985 onwards, the Commission began to impose 
swingeing anti-dumping duties on imported Japanese electronics 
goods, such as electronic typewriters and photocopiers. The Com- 
mission's actions set off a furious debate, with many economists 
and trade experts arguing that Europe w r as shooting itself in the 
foot by keeping out low cost goods. The reaction of the Japanese 
manufacturers was more pragmatic; they began to accelerate plans 
to build factories in Europe^ thereby hoping to avoid Europe's 
protective barrier. In 1987, Japanese inward investment into 
Europe almost doubled to $6,78 billion. 



270 



ALAN sugar: the amstrad story 

Alan Sugar looked askance at what was happening. Most of the 
anti-dumping complaints in consumer electronics were launched 
by the few established European giants, notably Philips of the 
Netherlands and Thomson of France. He believed that they 
wanted to build a wall around themselves to keep out competition. 
One result would be that prices to the consumer would remain 
higher than need be, reflecting the bloated overheads of com- 
panies like Philips. It would also be increasingly difficult for 
anyone to emulate entrepreneurs like himself, for a protective 
barrier around Europe would put a stranglehold on key electronic 
components. A budding Alan Sugar would either have to buy 
from a restricted set of high-cost suppliers or start making 
components himself- an almost impossible task for an individual 
entrepreneur, as Sugar explains: There's no such thing as a 
small chip plant or a small TV tube plant. It's either £300 
million or nothing.' 

Sugar publicized his arguments against the Commission's trade 
policies whenever he could, but he was too preoccupied to wage a 
campaign against it as he noted in 1986: The legislation designed 
to protect European industry is actually having the reverse effect. 
But I'm too busy running the business, and I have not the time, 
energy, patience or bureaucratic ways to fight the gnomes of 
Brussels.' 

Nevertheless, Amstrad's chairman could see the writing on the 
wall. He heard that an anti-dumping complaint about imported 
video recorders, spearheaded by Philips and Thomson, had been 
lodged with the European Commission in Brussels. Immediately 
he began to press Funai, the Japanese company which supplied 
most of Amstrad's VCRs, to set up a plant in Britain. In June 
1987, a new VCR factory was opened on Amstrad's Shoeburyness 
site; breaking new ground for Amstrad, it was a jointly owned 
operation - Funai had 51 per cent and Amstrad 49 per cent. The 
initial plan was to make 5,000 recorders a month, building up to a 
figure of 10,000. Three months later, in September, the European 
Commission duly launched its anti-dumping investigation of 
imported video recorders. 

Alan Sugar was finally forced to set out on what was, for him, 



•271 



DAVID THOMAS 

the unusual path of the political lobbyist. A new twist to the anti- 
dumping saga particularly concerned his company. The European 
Commission was planning to levy anti-dumping duties on the 
output of 'screwdriver plants' in Europe - assembly plants set up 
by Far East manufacturers which imported more than 60 per cent 
of their parts. Amstrad saw this provision as a threat to its new 
joint venture factory in Shoeburyness. 

In the hope that the British Government would put pressure on 
Brussels, Sugar lobbied Lord Young, then Trade and Industry 
Secretary. Amstrad was one of the first British companies directly 
affected by the Commission's anti-dumping measures. Young was 
on Amstrad's side, as he recalls: 

I had two or three meetings with him in the Department. And I 
had tremendous sympathy with him. I did all I could - he may 
not have thought I did - to help him through those problems. 
He had - and I know the type because I've been exactly the 
same way - no damn patience with all the officials, particularly 
in Brussels. But as a result of his case, I hope we're leaning 
more and more on Brussels. 

But the Government had no choice but to abide by the Euro-rules. 
In September 1988, the Commission imposed huge anti-dumping 
duties on five Japanese and South Korean companies. Funai's 
video recorders were saddled with an 18 per cent levy. Although 
Amstrad had ceased importing Funai VCRs in July 1988, the 
Commission, as expected, launched a 'screwdriver' investigation 
of Funai-Amstrad's Shoeburyness plant which eventually con- 
firmed the 18 per cent duty on its products. 

In his frustration, Alan Sugar made the supreme sacrifice: in 
December 1988 he travelled to Brussels to lobby the Eurocrats. It 
was not a meeting of minds. Having accused the Commission's 
officials of acting as the unpaid marketing department of Philips 
and Thomson, he returned disgusted with the ways of the Com- 
mission, describing their anti-dumping policies as '. . . a total 
sham. There's no logic in their decision. We were taking the 
markets because we're better marketeers, because we have a 



272- 



ALAN sugar: the amstrad story 

product that is good as far as specification is concerned and 
because we know how to advertise that product correctly.' 

The Amstrad chairman called on Mrs Thatcher to intervene: Tf 
she knew what was going on she would do her nut/ he told B.B.C. 
television. 

The only relief was a cut in the duty on Funai-Amstrad products 
on appeal to 13 per cent. But this still meant that Amstrad had to 
charge 13 per cent more for its video recorders than if the 
Commission had not intervened. Understandably, Sugar believed 
that such policies threatened the downward spiral in prices and 
the consequent rapid growth in demand which had been a feature 
of the VCR market in Europe until then. 

The simple Amstrad which Sugar had described in his City 
University speech was disappearing. Gone were the days when an 
ultra-lean organization in Britain could source as cheaply as 
possible in the Far East and then sell at minimum cost via agents 
throughout the world. But if the new Amstrad had to add to its 
previously tiny overheads and surrender a part of its old flexibility, 
Sugar's drive to increase the company's foreign sales also paid 
dividends. 

Britain accounted for only 42.5 per cent of Amstrad's sales in 
the financial year 1987-88. By building up its overseas sales, the 
company had correspondingly reduced its vulnerability to a down- 
turn in consumer spending in its home base. Amstrad was clearly 
stronger in some foreign markets than others, yet its overseas sales 
were broadly based: 19.4 per cent came from France; 17.3 per cent 
from Spain; 4.8 per cent from Germany; 4.3 per cent from Italy; 
4.9 per cent from other European countries; and 5.4 per cent from 
North and South America. It was an impressive basis on which to 
build. Only six years before, in 1982, Amstrad's overseas sales had 
amounted to £118,000. By 1988, they were running at £359.7 
million. 

Amstrad's new policy of establishing overseas subsidiaries had 
a dramatic impact on its once tiny head count. The work-force 
grew by two-thirds in 1987 and by a similar amount in 1988. By 
the end of financial year 1988-89, the company was employing 



273 



DAVID THOMAS 



over 1,600 people. In the most literal sense, Amstrad was no 
longer a one-man band. 



Continuing success in the computer market, particularly with the 
new top of the range PC 1640 personal computer and the PCW9512 
word processor, boosted Amstrad once more in the financial year 
1987-88. Profits increased 18.2 per cent to £160.4 million on sales 
up 22.2 per cent at £625.4 million. Sugar's warning in 1987 that 
Amstrad was facing a year of 'consolidation' while developing its 
overseas network had not prevented this continuing advance. The 
share price held up well in 1988 and by August Amstrad's stock 
was trading at 235p, which valued the company at more than £1 
billion. The value of Alan Sugar's own stake increased by £147 
million to £430 million over the course of 1988, making him the 
biggest stock market 'winner' of the year, according to a Sunday 
Times calculation. Amstrad had shrugged off the stock market 
crash of October 1987 as if it had never happened. 

In describing Sugar's immense wealth, newspapers tended to 
concentrate on the value of his holding in Amstrad, which 
fluctuated sharply with the Amstrad share price. But this was 
mainly a paper fortune, since he would never be able to realize it 
all. In many ways more significant was his real fortune and the 
opportunities it opened up for him. 

By 1986, Sugar had raised more than £34 million through 
successive sales of tranches of his Amstrad shares. His remaining 
stake in the company of over 40 per cent yielded large streams of 
earnings every year: in the financial year 1987-88, for example, 
his dividends amounted to about £3.5 million. Even if he had 
adopted the most passive and conservative investment policy 
imaginable - keeping his money in a building society - capital and 
dividend earnings on that scale would have generated huge 
additional income from interest. 

However, Sugar did not pursue a passive approach to his wealth. 
He had the normal spread of investments of the super-rich, with 
money invested at times in various overseas stock markets such as 
Tokyo; but he put the bulk of his fortune into commercial property 



274 



ALAN sugar: the amstrad story 

in Britain. It was relatively safe and, just as important, he enjoyed 
the buzz of managing his growing property portfolio. There were 
few things that relaxed Amstrad 's boss so much as closing a multi- 
million-pound property deal on the phone. The thing is I enjoy 
deals, and property is a dealing situation,' Sugar notes. 

Late in 1989, for example, he was becoming bored with a 
routine meeting in his office at Amstrad when one of his advisers 
called to discuss a building that he was on the point of buying: Ts 
the architecture O.K? ... No one's going to pull it down in a 
couple of weeks? . . . Lease is O.K? . . . Covenant sorted out? . . . 
Good, get on with it. That's it. Do the deal.' 

The person on the other end of the phone hesitated, but Sugar 
cut him short: 'Look, you're driving me mad. The deal's done, as 
far as I'm concerned. Get in touch with my solicitors. I want the 
contracts exchanged in an hour.' 

Sugar had bought another £2 million building which he had not 
seen. He would probably sell it again before seeing it. He employed 
his brother, Derek, to manage the day-to-day affairs of his property 
business such as paying the insurance and collecting the rents. 
'Everybody has a good laugh at me, because they can't believe the 
way I go about doing property deals. I buy and sell properties 
without even going to look at them,' he says. 

He has adopted a rule of thumb in his personal property 
dealings: not to bother with anything worth less than £2 million. 
'Some people might spend a week studying a property they're 
going to buy for £150,000. To them, £150,000 is a lot of money. 
But I'm used to dealing in millions. It sounds very blase to friends 
or associates or even advisers, but for me to spend £2 to £3 million 
is nothing. It sounds terrible, but it's true.' 

By the end of the 1980s, Sugar had also evolved more conven- 
tional ways to switch off from Amstrad. He had a circle of friends 
with whom he could relax, and he had mellowed from his younger 
days when his conversation had revolved obsessively around the 
progress of his small business. Often he does not even allude to 
some deal he has just completed at Amstrad when he meets his 
friends in the evening; they do not hear about it until they see it 
reported in the newspapers. 



275 



DAVID THOMAS 

Immediate neighbours, wealthy business people like the Sugars 
themselves, formed part of their circle. But most were friends they 
made when they first moved to Chigwell in the early 1970s. 
Increasingly wary of newcomers, in case they were seeking his 
company for ulterior purposes, Sugar felt most comfortable with 
his old gang of friends. 'I think he values his old friends too much 
to climb the social ladder so as to be out of reach. He doesn't have 
to be anything other than himself with us,' says Gerry Eriera, his 
long-standing friend. 

Many of the Sugars' friends inevitably inhabit different worlds. 
Barney and Myrna Lazarus, for instance, live in a small bungalow 
in Redbridge. A life-long cabbie, Barney has on occasion picked 
up Sugar or one of his business associates in central London. As 
Ann Sugar's first cousin, Myrna Lazarus was the link which 
brought the Sugar and Lazarus families together in the early 
1970s. 

A meal in a local Italian or Chinese restaurant still ranks as an 
evening out for the Sugars, and the Lazarus couple admire the fact 
that their tastes and lifestyle have not radically changed as their 
wealth has mounted. They also respect Alan Sugar for not 
patronizing them with his wealth. On special occasions, the Sugars 
may invite their friends out for a meal, but Sugar does not make 
people like Barney and Myrna feel uncomfortable by flashing his 
money around: 'What I like about him is that when it comes to 
paying the bill, he'll only pay his share. He doesn't try and take 
over and say, "I'll pay for it all." I admire him for that,' says 
Barney Lazarus. 

Most winters the Sugars visit their holiday home in Florida. A 
four-bedroomed house with a small patio, a swimming pool and 
its own tennis court, it backs on to the eleventh hole of a golf 
course in Boca Raton, the town 35 miles north of Miami where, 
ironically, I.B.M. built its first personal computer. Some of the 
Sugars' wealthier friends, like the Erieras, have also bought a villa 
in the same complex. Others, like the Lazaruses, come over to 
stay. As a result, a party of Sugar's friends tends to fly over most 
Christmases, joining the 'snowbirds', the New Yorkers who 
migrate to Florida for the winter. 



•276 



ALAN sugar: the amstrad story 

Sugar tries to keep his workload down to a minimum during his 
trips to Florida, spending perhaps an hour a day in his small office 
there. Otherwise, he plays tennis, jogs, rides his bike and relaxes 
with his friends. Stories abound about his sense of fun, whether 
mimicking American accents or impersonating Charles Aznavour 
from the little balcony overlooking the large living room in his 
house. 

Sugar's time away from work revolves around his family. He 
has never been tempted into the kind of jet-setting existence which 
could dilute this fundamental commitment. He is proud that his 
children - two sons and a daughter - were not spoilt by his great 
wealth. They were educated at local day schools, mostly in the 
private sector because Sugar feels that state schools have deterio- 
rated since the days when he attended Brookhouse. Pocket money 
was limited when they were children, and as young adults they 
were given modest cars. 

Amstrad's chairman hopes that his two sons will follow him in 
the business, but he insists that they will have to earn promotion. 
Simon, the elder, did a stint after leaving school working for 
Dixons in order to understand the viewpoint of Amstrad's cus- 
tomers. Dixons laid on a programme for him, so that he could see 
different facets of the business, but he was given no special 
treatment when serving in a Dixons store in the West End of 
London. Stanley Kalms, Dixons chairman, says: 'He wanted to 
learn retailing and he did. He is a very quiet, hardworking and 
modest lad. He was treated just like anyone else - well, just like 
anyone else who had Alan Sugar for a father. I'm sure he was 
treated nicer than he should have been, but it wasn't company 
policy.' 

The younger son, Daniel, went to work for Amstrad as an 
assistant in the marketing department as a 16-year-old straight 
from school. There was no hint of a 'one day, son, this will be all 
yours' attitude in the way Sugar treated his son. On the contrary, 
on Daniel's first day in 1987 Sugar dragged him in to see Thomas 
Power - Amstrad's marketing manager at the time - and said, 
'This is Daniel. This is my son. I want you to teach him all that 
marketing crap you lot go on about down here. I want you to 



277 



DAVID THOMAS 

teach him everything you know. If he gives you any talkback, send 
him up to me. And if he doesn't do what you tell him, throw him 
out the door.' 

Daniel was used to his father's jokey-aggressive manner. Power 
found him to be a modest, well-balanced lad, who had coped well 
with the problem which confronts many children of the successful 
- living in the parental shadow. Yet Power also detected an 
entrepreneurial spark in Daniel and says: 'He's got the bolshy, 
"Oh yeah, I'm a trader, I'm a banana salesman and I'll sell you a 
satellite dish on the side," attitude. He's got all that. He has a fast 
mind in the trading, not in the academic, sense.' 

As far as Sugar's own parents were concerned, it was an uphill 
struggle persuading Fay and Nathan to adapt to his great wealth 
and to let him spend money on them. It took him years to convince 
them to move out of their flat in Hackney, which was rapidly 
succumbing to the worst forms of inner city blight. He wanted to 
buy them a flat in Redbridge, but initially they resisted the move 
as too extravagant. He paid for them to take holidays in the 
United States and Israel, but had to tell white lies about the first- 
class air tickets he bought. Sugar said they had been given to him 
in the course of business, since without this subterfuge his parents 
would have worried about the expense. 

Sugar's natural competitiveness did spill over into one area of 
his private life - his tennis - which absorbed more and more of 
his time after he gave up flying in the early 1980s. The game 
brought Sugar's business and private worlds together one evening 
in June 1989, when Amstrad sponsored the pre-Wimbledon pro- 
celebrity tennis tournament at the Royal Albert Hall. Television 
personalities, actors and sportsmen like Terry Wogan, Jimmy 
Tarbuck, Dennis Waterman and Bobby Moore, were among the 
celebrities who turned out to support the event. Amstrad's 
business associates joined a large contingent of Sugar's friends, 
who became more lively as Sugar and his partner - Indian tennis 
star Vijay Amritraj - progressed through the early rounds. 
Banners supporting Sugar were unfurled from the balcony of the 
Royal Albert Hall, while shouts of 'Come on, Alan' echoed round 
the arena. 



278 



ALAN sugar: the amstrad story 

The most interesting game was the semi-final which pitched 
Richard Branson against Alan Sugar. The conduct of two of 
Britain's best-known entrepreneurs on the tennis court spoke 
volumes about their contrasting business styles. Virgin's chairman 
was the perennial public schoolboy, jokily hitting the tennis ball 
as high into the reaches of the Albert Hall as possible, yet still 
managing to play well enough to keep Sugar on his toes. Amstrad's 
boss was the gritty East Ender, focusing on the task of winning the 
match, though occasionally finding time to flash a sheepish grin at 
his supporters. Sugar won the day on that occasion and the Sugar- 
Amritraj duo went on to win the final, giving his friends an excuse 
to celebrate - not that they seemed to need one. 

One of several charitable activities which Sugar undertook in 
the second half of the 1980s, the tennis evening raised some 
£170,000 for the Muscular Dystrophy Group. In 1986 Sugar 
created a personal foundation to distribute money to charity. The 
Alan Sugar Foundation has given funds to local causes, the largest 
being a donation of almost £1 million to build a new wing for an 
old people's home in Ilford. It has also donated to national 
charities such as the Great Ormond Street children's hospital 
appeal. Sugar never broadcast these activities and even his close 
friends did not know about his foundation until some time after it 
was created. 

Amstrad's chairman has worked for Business in the Community 
(BiC), a group which channels business help towards the inner 
cities. One day Sugar went with the Prince of Wales, BiC's 
president, to Hartlepool - an unemployment black spot in the 
North of England. Hartlepool reminded him a little of the East 
End of his childhood. It seemed to be full of people like his father, 
who lacked the confidence to realize they could branch out on 
their own. Sugar, who was there to talk to a group of unemployed 
youngsters, told them, 'They could become gardeners, window 
cleaners, painters and decorators. Perhaps some of them might 
even be able to employ one or two other people. I suggested they 
should have the confidence to think of starting up their own 
business.' 

As Sugar became more accustomed to appearing on platforms 



279 



DAVID THOMAS 

and before the television cameras, he revealed an untutored talent 
for public speaking. Business people are often intensely dull 
speech-makers, but he has a natural sharpness and sense of 
humour in public. Jimmy Tarbuck, the Liverpool comedian who 
has hosted several Amstrad business events, has watched Sugar's 
confidence in public grow and admires his timing: 'I'd have no 
complaints about working on a script with him. He'd be very good 
to bounce off. His delivery is good enough and he'd be all right 
with a line,' Tarbuck says. 

Yet Sugar's facility in public, plus his hard-hitting language, 
disguises a deep-rooted shyness which surprises people who know 
only his reputation for bluster. This diffidence helps to explain his 
indifference to personal publicity. Interest in Sugar and Amstrad 
exploded after the success of the PCW8256 word processor. Nick 
Hewer, Sugar's public relations adviser, has regularly to decline 
forty or so requests a month for interviews, guest appearances and 
talks by Amstrad's chairman. Sugar has turned down invitations 
for which many people wait a lifetime - to appear on the Desert 
Island Discs radio show, or as one of the main commentators on 
the B.B.C.'s post-Budget programme. Hewer had to struggle hard 
to persuade him to be profiled or interviewed by some of the 
world's most prestigious publications like Time, Fortune and the 
Wall Street Journal. The only way Hewer could secure Sugar's 
agreement to a profile in Time was by showing him the rate for a 
page advert in the magazine and pointing out how much he would 
save Amstrad by agreeing to the interview. 

Sugar was being drawn - in many ways against his natural 
inclination - on to a larger stage. In 1988 he was asked by Lord 
Young, the Trade and Industry Secretary who later became 
deputy chairman of the Conservative Party, to spearhead the 
Government's 1992 advertising campaign. Young chose a handful 
of well-known business leaders to spread awareness in the business 
world of the European Community's programme to remove inter- 
nal trade barriers. The Minister wanted people who would be 
widely recognized on television and on the advertisement hoard- 
ings - people like Sugar, Richard Branson, Sir John Harvey-Jones 
(former head of LCI.) and Sir John Egan, who led Jaguar out of 



280 



ALAN sugar: the amstrad story 

public ownership. 'All of them were people who stood out above 
the crowd. And there's no question that Alan Sugar was the 
foremost British leader in the computer industry and in electronics 
goods generally,' says Young. 

Disliking the script written for his television commercial by the 
Government's advisers, Amstrad's chairman phoned Young and 
said he could do better himself. Young readily agreed: 'I thought 
that since he knew a thing or two about marketing, I wouldn't 
worry.' 

Sugar's 1992 advert was simplicity itself. It showed him stand- 
ing in front of Amstrad computers marked for Britain, Germany, 
France, Spain and Italy. At present, he told the camera, Amstrad's 
computers have to be adapted for national markets. After 1992, 
they will all be the same. Research carried out for the Trade and 
Industry Department showed that this advert was one of the most 
effective in arousing business people's interest in 1992. 

Invitations to lunch with the Queen at Buckingham Palace and 
to seminars and dinners at No. 10 Downing Street also landed on 
Sugar's desk. Yet he did not always seem at ease in the great 
world of politics, at times giving the impression that he would 
rather be running his business. Lord Young, who first met him at 
a Downing Street seminar, believes this may reflect the way 
Sugar's business is developing: 

I suppose during the early years of the decade, up to 1986 or 
1987, Sugar was riding a tiger which was almost impossible to 
control. Then, just when it's established and he's worth several 
hundred million pounds, Amstrad develops some problems. 
He's got to overcome those. But by the early 1990s, the problems 
could be behind him and he's got a steady business which is 
growing fast. It may come to the stage when it's not a 24-hour- 
a-day job for him to run his business. Then he may look for 
other things to do. It may be charitable or political activities. 

Young also felt that Sugar's unease in governmental circles flowed 
from his outsider status - his history of having clawed his way up 
from the East End. 'He's still, I think, got a slight air of insecurity 



281 



DAVID THOMAS 

about him. He hasn't shown any sign of wanting to come into the 
fold. That may happen later on, perhaps when he's Sir Alan Sugar 
or Lord Sugar of Chigwell, or whatever else happens in their 
dotage to great industrialists. Who can tell? He may not. He's 
never struck me as the sort of person who sets great store on being 
a member of the Establishment.' 



282 



• 15- 
Pennies from Heaven 



Once Alan Sugar became a public figure, he was inevitably the 
target for unwanted attention. Amstrad headquarters began to 
receive phone calls from cranks, people on the make and those 
with nothing better to do than call up someone they had read 
about in the newspapers. Many of these stray callers claimed some 
long-lost association with Sugar or his family. As Sugar describes 
the problem: 

The phone does not stop ringing from nine o'clock in the 
morning to six o'clock at night. That school I went to in 
Hackney must have had 250 million boys in it, because every- 
body that phones claims to know me, or went to school with me, 
or knows a brother who's got an uncle who knows an aunt who 
knows me. Honestly, I can't talk to everybody who picks up the 
phone to me every day. So Frances, my secretary, has got this 
patter of politely saying to people, 'Who are you? Do you really 
know Mr Sugar?' 

One day in May 1988 the routine backfired. Frances, who is 
politeness personified, came into Sugar's office and said, 'Mr 
Rupert Murdoch wants to speak to you on the phone.' 

Sugar kept his eyes firmly on the pile of papers on his desk. 
'Never heard of him. Tell him to piss off, clear off, I'm not 
interested. I bet he thinks he went to my school.' 

Amstrad's chairman explains this surprising reaction: 



283 



DAVID THOMAS 

The problem with me is I'm very bad on names and also I'm 
very bad at reading the newspapers. I'm very unworldly in the 
sense that outside the electronics and computer industries, I 
don't know the names of big-shot businessmen. If you asked me 
who is the chairman of B.A.T. or Shell or British Telecom, I 
wouldn't know. There are people who go around name-dropping 
all the time, saying Sir John This or Sir John That, and you're 
supposed to know who they're talking about. I really don't. 
Sometimes people think I'm lying, but I'm not. 

Sugar could sense Frances hesitating, so he said, 'Who the 
bleeding hell is this Murdoch? Go and find out what it's about.' 

Frances returned. 'The lady says Mr Murdoch doesn't normally 
tell people what his business is before speaking to them himself. 
She rang off, but I have taken their number.' 

Sugar lifted one eye from his task. 'Well, who is this bloke 
Murdoch?' 

Frances replied: 'He owns the Sun, The Times, the Sunday Times, 
the News of the World and Today newspapers and he's the one who 
had the big confrontation with the unions.' 

'Oh my God,' said Sugar. 'Quick, get him back on the line.' 



Alan Sugar was one of the first businessmen in Britain to grasp the 
potential of a new method of delivering television - direct broadcast- 
ing by satellite. Technological advances which allowed programmes 
to be beamed down by satellite, coupled with a political climate in 
Britain which favoured more competition for the established tele- 
vision channels, combined to usher in this broadcasting revolution. 

In 1986 Amstrad's chairman was visiting the headquarters of 
Granada, the diversified group whose interests include Granada 
Television and high street rental and retail chains specializing in 
electrical goods. He was there to discuss Amstrad's business as a 
supplier to Granada's shops, but the conversation roamed more 
widely. Granada's executives told him that they were forming a 
consortium to bid for Britain's first satellite broadcasting franchise. 
The plan was to provide three new channels and they wanted to 



284- 



ALAN sugar: the amstrad story 

know whether Sugar would like to join the consortium. The entry 
ticket would be an initial £10 million equity stake. Amstrad's role 
would be to supply a crucial ingredient in the satellite broadcasting 
mix - the dishes that receive the signals from the satellite. 

Sugar liked the proposal. He believed Britain to be seriously 
under-provided with television channels compared with the 
United States, where the average American can receive almost 
fifty stations. He felt there was a healthy appetite for more 
television among the people who bought Amstrad's audio and 
video products - the original truck-driver-and-his-wife Amstrad 
customers who had no qualms about the amount of television they 
watched, unlike the chattering classes who dominated London's 
opinion-forming circles. After all, Sugar was happy to admit that 
he relaxed by watching 'American garbage' on television. 'Satellite 
broadcasting is good for our type of British consumer,' he was 
later to tell a meeting of City analysts. 'It may not appeal to you 
lot, but the Sun readership and the News of the World readership are 
going to buy that stuff.' 

Sugar agreed to join Granada's consortium, which called itself 
British Satellite Broadcasting (B.S.B.). It boasted an impressive 
line-up: Virgin, the music group built up by the ever-youthful 
Richard Branson; Pearson, the blue-blooded blue chip company 
whose many interests include the Financial Times; and Anglia 
Television, together with Granada and Amstrad. It appeared to 
be a well-balanced team, with the television expertise of Granada 
and Anglia complemented by Pearson's heavyweight respectability 
and the marketing flair of Amstrad and Virgin. With Amstrad's 
participation, B.S.B. could point to a credible source of dishes for 
its satellite service. 

The Independent Broadcasting Authority, the television regu- 
latory body, had to judge between the bidders for Britain's first 
satellite broadcasting franchise. At 3 pm on 11 December 1986, 
representatives of the five consortia bidding for the franchise were 
summoned to the I.B.A.'s offices and asked to wait in separate 
rooms until handed an envelope by Lord Thomson, LB. A. chair- 
man. When the five opened their envelopes, they found that the 
I.B.A. had also been impressed by B.S.B.'s mix of expertise. B.S.B. 



285 



DAVID THOMAS 

was the winner and Amstrad seemed to have found itself a new 
line of business. 

But it was not to be. Amstrad quickly began to cool to its new 
partners. Views differ as to why this happened. A senior B.S.B. 
executive of the time voices what was a common perception of the 
reason for Amstrad's disenchantment with B.S.B: 

Amstrad came to realize that it was not going to get a monopoly 
over the supply of receivers to B.S.B. — which perhaps it had 
thought it might. Amstrad also perhaps thought that it might 
get earlier or privileged access to information about things like 
the specification. But B.S.B. insisted on dealing with Amstrad 
like any other supplier. 

Sugar rejects this version of events: 'There was never any question 
of us going to get any exclusive rights or any preferential rights, or 
any privileged knowledge of the receiver.' 

Amstrad's chairman explains that he rapidly became uneasy 
with the B.S.B. operation. Jim Rice, the company's representative 
at the B.S.B. 's partners meeting, regularly reported back to him 
on progress, and once or twice Sugar went himself to check out 
developments. His dissatisfaction mounted during the early 
months of 1987 as the moment approached when the founding 
partners would be asked to put up their cash: 

You get a feeling about people and where their priorities lie. 
When we started listening to them, it seemed their priorities lay 
in the size of the office suites they had to have and all that kind 
of stuff. Their own little luxuries came first. I was being asked 
to entrust my money to a group of people who were going to run 
this show. And there was no way I felt comfortable in putting 
my money in the hands of those people. I looked at the 
extravagance of what they were planning to do there and when 
it was time to part with the old doshoroney, I said no thank you 
very much. 

B.S.B. looked set to make considerable cash demands on its 
partners for several years before its service began to operate. 



286- 



ALAN sugar: the amstrad story 

Continuing participation would have cost Amstrad significant 
amounts with no immediate return. Developing equipment for 
B.S.B, would be a long and expensive process, because B.S.B. was 
committed to using a new broadcast transmission system called 
MAC - eventually it opted for the D-MAC version. D-MAC will 
ultimately deliver a better picture than the older PAL system 
currently used in British broadcasting, but its superior picture is 
at present available on only a relatively small number of new high- 
tech screens. On the great bulk of sets there is little to choose 
between the quality of pictures transmitted through the two 
systems. Backing the development of D-MAC equipment seemed 
to Sugar a costly irrelevance. 

On 15 May 1987 Amstrad quit B.S.B. The news was buried in 
the simultaneous announcement that B.S.B. had provisionally 
arranged another £200 million of funding and attracted a batch of 
new investors. Sugar recalls a conversation with the chairman of 
Virgin, another B.S.B. partner, at the time: 'I remember Richard 
Branson said to me, "I hope you're wrong, because I'm staying 
in." I said, "Well, maybe it's more up your alley than mine." And 
as we all know, Virgin later got its money out of Bondy and ran 
for the hills.' 

In 1988 Virgin sold its B.S.B. stake to Alan Bond, the 
Australian businessman. Thus B.S.B. had lost the two people - 
Branson and Sugar - who gave its initial line-up entrepreneurial 
fizz. 

Amstrad did not entirely lose touch with B.S.B., which in early 
1988 was searching for possible suppliers of receiving equipment. 
As many as a hundred suppliers were considered; a short-list of 
fifteen, including Amstrad, was announced in early May 1988. Yet 
Amstrad's heart appeared not to be in it. The company did not 
ask B.S.B. for the full documentation needed by potential sup- 
pliers, which was unusual for such an aggressively competitive 
firm as Amstrad. Unbeknown to B.S.B., Alan Sugar had received 
Rupert Murdoch's phone call shortly after B.S.B.'s short-list had 
been published. 

Murdoch's News Internationa! had been a member of one of 
the consortia which failed to win the LB.A.'s satellite broadcasting 



287 



DAVID THOMAS 

franchise, but Murdoch refused to regard this as the end of the 
story. During 1987, while B.S.B. was making its plans in the full 
glare of publicity, he was quietly looking for another way into 
satellite broadcasting. It became apparent that Astra, a private 
sector satellite service based in Luxembourg, offered a route. Astra 
was in the business of launching satellites capable of transmitting 
television from the Ariane space rockets which blast off from 
French Guiana. The question was: would it make commercial 
sense for Murdoch to book space on an Astra satellite for a satellite 
television service of his own? 

The small team of executives researching the project for Rupert 
Murdoch quickly identified the critical factor. Peter Smith, a 
softly spoken Australian engineer who was later to become Sky 
Television's director of operations and planning, remembers: T'd 
done some financial projections for Sky and it turned out that 
what is absolutely crucial is the rate of growth of receivers in the 
market place. The profitability and survival of the organization 
is far more sensitive to that than to virtually every other factor 
put together.' 

Murdoch concluded that he would need a supplier who could 
produce large volumes of receivers at rock-bottom prices. He and 
his team identified £200 as the magic figure below which receiving 
equipment might sell in great numbers. Smith, nothing if not 
methodical, urged on Murdoch the need for market research to 
see if people would buy receivers at that price. Murdoch disagreed: 
'You cannot do effective market research on something that 
nobody has ever had, because people do not know enough about 
it to be able to say whether they would buy it.' 

Murdoch and Smith were still debating the merits of market 
research while scouring the Far East for sources of receivers. One 
of their appointments was with Akio Morita, the legendary 
Japanese engineer-businessman who built up Sony to become one 
of the world's great consumer electronics innovators. Murdoch 
and Morita had met a few times, so they greeted one another 
fulsomely, took the lift in Sony's Tokyo headquarters to the top 
floor and waded through the carpet to Morita's office. Following 

•288- 



ALAN sugar: the amstrad story 

the protocol for meetings between corporate barons in Japan, 
they drank tea and made speeches praising each other's company. 

During the course of this ritual, Morita asked Murdoch, 'Do 
you ever have trouble with your marketing people?' 

Murdoch looked at Smith and replied, 'Yes, as a matter of fact, 
I do.' 

Morita: 'Do they always want you to do market surveys?' 

Murdoch: 'Well, yes, why though do you say that?' 

Morita: 'I had an idea for my company once. I thought of a new 
product. I wanted to build a tiny little thing which played cassettes 
and which people could carry in their pockets while they were 
jogging or working. But my company told me it would be terribly 
expensive to build, so I had to do a market survey. But I had a 
gut feeling that it was going to be right.' 

'Gut feeling' is one of the favourite expressions of Rupert 
Murdoch, who was beginning to enjoy the story immensely. 

Morita continued, 'My company then asked me what I wanted 
to call this machine. I said, "A Walkman." And they told me, 
"That's a terrible name. We'll have to do a market survey on it.'" 

The proposal for a market study of satellite broadcasting's 
potential was not raised again within News International, but 
Murdoch's trip to the Far East was less fruitful in other respects. 
He and Peter Smith did the rounds of the great names in Japanese 
electronics - Sony, Toshiba, Mitsubishi and so on. Even though 
Japan already had a satellite broadcasting system, none could 
offer Murdoch what he wanted; the cheapest dish any would 
promise to produce would have cost £300 - 50 per cent above 
Murdoch's target price. Also they wanted plenty of time to develop 
the equipment and precise information from Murdoch about the 
market - impossible questions to answer since the market did not 
exist. 'The Japanese manufacturers turned out to be very conserv- 
ative. They were very European in their approach,' Smith says, 
looking back. 

Murdoch continued to draw up plans for Sky Television, but 
the lynchpin problem still had not been solved, as he explains: T 
felt the whole thing was contingent on getting a supplier of 
receiving equipment at a low price. If we waited for the Japanese 



•289 



DAVID THOMAS 

or the French to come late and expensively, we'd go broke waiting 
for the market to develop.' 

In May 1988, it occurred to Murdoch that he need not look to 
Japan or the Continent for cheap receiving equipment. The answer 
lay much nearer to News International's home in the East End of 
London. In everything he had done, Alan Sugar had shown that 
he could produce large volumes cheaply. Why not ask him to 
make Sky's receivers? 

By the time Murdoch phoned Sugar, he had already worked out 
the main parameters of Sky Television: he knew the price he 
wanted for the equipment, and he thought it sensible to broadcast 
using PAL rather than D-MAC. But in his first conversation with 
Sugar, News International's boss wanted to take the measure of 
the man. So he sounded him out on these issues. 

Sugar's answers to Murdoch's questions were to the point: 

The success of a satellite depends on a triangle. The triangle is 
like this: the satellite, good programmes and receiving equip- 
ment. The programmes have got to be good and that means it 
must have a movie channel. It will be a failure without a movie 
channel. The equipment's got to be cheap. And we can transmit 
in PAL, because that's all that's available now. If anyone tells 
you anything different, it's a load of cobblers because D-MAC 
does not exist. There are no D-MAC chip sets. All this quality 
cobblers that B.S.B. is talking about through using D-MAC is 
also a load of bloody rubbish, because the quality enhancement 
from a D-MAC transmission will only be realizable when things 
like high definition television come on the market. In the 
meantime, there are 22 million televisions out there which can 
only receive in PAL, and there's nothing which anybody can do 
to improve the picture. That is a fact of life. If something is 
transmitted in D-MAC, it's got to come down here, be translated 
into PAL and be pumped into your television. Otherwise you 
can't receive it. 

So convinced was Sugar of this point that in June 1988, after his 
link with Murdoch had been made public, he issued a challenge. 



290 



ALAN sugar: the amstrad story 

Pointing to the television in his office, Sugar said, 'I'll offer £1 
million to anyone who can make a better picture than that. D- 
MAC television sets haven't even been developed yet.' 

B.S.B. took up the challenge, showing how a conventional 
television equipped with a special adaptor could offer a better 
picture by receiving D-MAC signals. In October, B.S.B. ran full- 
page advertisements proclaiming, 'Dear Alan, Looks like your 
PAL just cost you a million.' The adverts asked him to send 
£1 million to a charity of his choice. 

But Sugar would not play ball. He accused B.S.B. of ignoring 
the terms of his challenge, which applied to the television set in 
his office - similar to that in 95 per cent of British homes. T am 
ready to stand up and be counted, proved wrong, eat my hat and 
part with £1 million, if I am wrong, 5 he said. 

The phone call between Murdoch and Sugar showed that they 
were thinking along the same lines, but Murdoch did not take the 
conversation further immediately. He wanted to check Sugar out. 
After all, Amstrad would be able to break Sky if it failed to deliver 
the receivers. Murdoch telephoned some contacts in the top 
echelons of British business, including Lord Weinstock, to ask 
about Sugar. He liked what he heard and called the Amstrad 
chairman again after a couple of days. This time Murdoch made 
a proposal: Would Amstrad be able to make very large numbers 
of receivers to retail at £199 each? Sugar replied he would find out 
and let him know in a matter of days. 

Amstrad's chairman told Bob Watkins, his technical director, 
that he had a day or two to find out whether Amstrad could make 
a profit out of receivers selling for £199 in the shops. He briefed 
Watkins to be flexible: 'There must be some dustbin lid manufac- 
turer who can make the dish for us, or somebody up in the 
Midlands whose business is waning because the car industry is 
going down the pan. See if we can do it.' 

Bob Watkins has become used to being given tasks like this by 
Sugar. As a result he has developed an unusual blend of 
engineering and commercial nous. He often has to decide whether 
Amstrad can make something for a certain price to a deadline 
which would give a more conventional engineer nightmares. 



291 



DAVID THOMAS 



Sugar's overriding need for speedy decision-making has de- 
manded it. And Sugar says that Watkins's calculated gambles 
have proved right within plus or minus 5 per cent 9V2 times out 
of 10. 

Watkins gave Sugar the news a couple of days later. Amstrad's 
technical director was still waiting for a few more quotations to 
come in, but his gut feeling was that it could be done. This time 
Sugar phoned Murdoch: 'We'll take a gamble and I'll say, yes, we 
can do it for £199. But what we will do, is have one model at £199 
and, as a contingency plan, another model at £259 with remote 
control, which gives us a little bit more margin in case we've made 
a mistake.' 

Murdoch agreed to this proposal. Now it was Sugar's turn to 
make a few demands, as he explains: 'I put my humble hat on and 
said, "Look, I'm only a little humble lad from Brentwood. I 
haven't got the multi-mega-millions of your organization. You're 
not asking me to ponce around making 5,000 pieces; you want me 
to put hundreds of thousands of receivers out there. So what 
happens if the bleeding rocket doesn't go up?'" 

Sugar would have to start work on the receivers immediately, in 
May 1988, knowing that two critical dates lay in the future. One 
would be later in the year when an Ariane rocket would send the 
Astra satellite into orbit. However, Ariane spaceships had been 
known to fail to leave the launch-pad or explode on take-off. The 
other critical date would be in February 1989, when Sky was due 
to start broadcasting; if Sky pumped out feeble programmes, then 
Amstrad could be left with large numbers of unwanted dishes. 

Alan Sugar and Rupert Murdoch met to hammer out their final 
agreement. Sugar hit it off with Murdoch, just as he had with 
Weinstock, recognizing a kindred spirit - someone also used to 
making swift decisions, taking risks and getting his own way. It 
did not take the two entrepreneurs long to agree terms. 

For his part, Sugar agreed to set the wheels in motion to make 
a million satellite dishes in a year. He understood from the start 
that it was a non-exclusive relationship: other suppliers would be 
free to make receivers for Sky. At this stage, Murdoch did not 
place an order for a single dish with Sugar. His part of the deal 



292 



ALAN sugar: the amstrad story 

was to agree to repay all Amstrad's costs in one of three cases: 
either the rocket failed to go up; or the satellite failed to make 
satisfactory test transmissions; or Sky failed to go on air with four 
channels. As Sugar recalls, 

He indemnified us that he would go ahead and start with four 
channels. So we wouldn't go to the trouble of getting a million 
satellite dishes organized and all he was going to put on the air 
was some tenth rated Channel 4. Sky would definitely have a 
sports channel, a movie channel, a light entertainment channel 
and a news channel. We went on our own gut feeling and the 
indemnification. 

The Murdoch-Sugar link-up, though not its precise terms, was 
announced on 8 June 1988, barely three weeks after Murdoch's 
first phone call to Sugar. Amstrad's chairman had been busy on 
other fronts in this short space of time, and he was able to tell the 
assembled journalists that Dixons had agreed to take 500,000 
dishes from him. The Murdoch-Sugar press conference was as 
crowded as the Sugar-Sinclair event of two years before; it was 
before the assembled throng that Murdoch dubbed Sugar 'prob- 
ably Britain's greatest entrepreneur'. 

Sugar busied himself tying up deals with the contractors which 
would make the various parts of Amstrad's receiving equipment. 
The order for the convex metal work of the dish was placed with a 
Birmingham-based metal-basher, as Sugar had anticipated - 
although the supplier, Concentric, was not known as a dustbin-lid 
manufacturer and was enjoying booming sales to the European 
motor industry. The electronics for the receiver, a black box w r hich 
sits on top of the television, was made in the Far East. This left 
just one main part for the Amstrad dish: the low-noise block 
converters (LNBs), which process the signals transmitted by 
satellites and which stand out from the middle of the metal dishes. 

Amstrad was combing the Far East for suppliers of the LNBs 
when Sugar became aware that Britain's General Electric Company 
was interested in the work. The interest flowed directly from Lord 
Weinstock, G.E.C.'s managing director, who had kept in touch with 



293 



DAVID THOMAS 



Sugar since the abortive talks about a joint venture between their 
two companies. Weinstock had set up a research team to look at 
ways of generating spin-off consumer products from the company's 
work in defence electronics, an area where G.E.C. is immensely 
strong. G.E.C.'s managing director reckoned that its expertise in 
radar and similar fields could be brought to bear on the LNBs, 
which Amstrad would need in huge numbers. 

Bob Watkins negotiated the outline of a deal with his counter- 
part at G.E.C. but the talks became stalled over price. It looked 
as though G.E.C. would lose Amstrad's order to a Japanese rival 
and Sugar decided to intervene: 'It had to be buttoned up and 
sometimes there's only one way of sorting things out. So I decided 
to go straight to Arnold. I told him what was going on and that 
G.E.C. was about to lose the order,' Sugar says. 

After some haggling on the phone, Alan Sugar and Arnold 
Weinstock closed the deal there and then. Sugar formed the 
impression that Weinstock enjoyed the deal-making process as 
much as he did: 'Arnold was excited by it. He liked doing the deal. 
It reminded him of his early days, because him and me were 
haggling on the phone over 50 pence, which is probably something 
he hasn't done for 30 years.' 

The collaboration with G.E.C. over the LNBs played an import- 
ant part in the evolution of Sugar's thinking. The growth of 
protectionist trade sentiment in the European Commission, cou- 
pled with rising wages in the Far East, was forcing him to consider 
moving more of Amstrad's manufacturing back to Europe. It was 
a change which he contemplated reluctantly, since his experience 
of British manufacturing had been largely negative. But G.E.C.'s 
work on the satellite contract began to give him a new perspective. 
'Arnold pulled together a team of people and instructed them to 
get on with it. And they did. I must give them credit where credit's 
due. It brought back a lot of my confidence in British electronics 
when I saw them do that.' 

Sugar's relations with Murdoch settled into a groove. The two 
men tend to talk on the phone every few weeks, not just to sort out 
any immediate problems but also - once Sky Television was 
launched - to discuss how the service is faring. Murdoch says he 



294 



ALAN sugar: the amstrad story 

rates Sugar's advice highly because he trusts his understanding of 
the ordinary British consumer: 'I call him once or twice a month 
to see how things are going - to keep in touch and get his opinion 
on the market place,' Murdoch explains. 

In the early days, Sugar viewed the newspapermen who ran News 
International as having strayed into a line of business - consumer 
electronics - of which they were profoundly ignorant. As a result, he 
tended to give short shrift to Murdoch's latest query or anxiety, a 
pattern established at their first meeting in May 1988: 

Murdoch: 'What do you think of this B.S.B.? They claim they're 
going to be on the air in September 1989.' 

Sugar: 'No way.' 

Murdoch: 'Well, they say they're going to do it.' 

Sugar: 'Well, I say they're not.' 

Murdoch: 'Why not?' 

Sugar: 'Because they haven't got any chips.' 

A few months later, after the Sugar-Murdoch deal had been 
struck, Murdoch warned Amstrad's boss that other manufacturers 
were now planning to make receivers for Sky. 

Murdoch: 'Sinclair is going to make some dishes for us. Toshi- 
ba's also coming along and will be delivering dishes.' 

Sugar: 'They won't turn up. They won't deliver you any by the 
date they say, and half of them won't work.' 

Sugar began to worry about the view Murdoch was forming of 
him: 

The kind of impression he must have got of me in the beginning 
was: this geezer's a bloody great bighead. He knows everything. 
He says everything quickly. He's so convinced and incisive. But 
the point is they were trying to tell me about my business, the 
electronics business. They were unqualified people talking about 
things they didn't know anything about. They probably thought 
everything I was saying was just salesman's talk. But it's all 
come to pass. 

Sky and Amstrad had to resolve several problems before Sky's 
launch, the most serious of which concerned Sky's plans for a 



295 



DAVID THOMAS 

movie channel. Originally, Murdoch had intended Sky's movies 
to be free to all Sky viewers, but he changed his mind during the 
course of 1988. He realized that he could not buy the rights to 
many movies in European countries outside the U.K. and Ireland, 
because they were already owned by various parties who had no 
intention of selling them. Sky would have to put a scrambler on its 
system to prevent people on the Continent watching its movie 
channel in order to avoid actions for breach of copyright. But, 
Murdoch concluded, in that case why not turn the necessity to 
scramble into a source of revenue? 

Only those viewers in the U.K. and Ireland who paid a monthly 
subscription would have access to a descrambler to see the movie 
channel. In this way, Sky would gain an additional stream of 
cash, which would be particularly valuable while its viewing 
numbers - and hence attractiveness to advertisers - were building 
up. Sky would need this revenue badly, since competition 
between itself and B.S.B. for programming rights pushed up costs 
sharply during 1988. B.S.B. committed more than $700 million 
over five years for the rights of five U.S. studios. 'When we came 
to negotiate for the movies, we found B.S.B. already there 
spraying literally hundreds of millions of dollars around Holly- 
wood,' Murdoch says. 

Sky considered two main descrambling methods. The subscriber 
could be given a smart card, which would fit into the decoder; or 
a signal could be transmitted remotely to the decoder. At first 
Murdoch was set against the smart card, believing it to be an 
expensive solution. By contrast, Sugar championed this idea to the 
point of lining up a sub-contractor who could make them for 
Amstrad. He doubted whether remote descrambling was available 
except in a form which would be easy for hackers to decode, and 
mounted a sustained campaign to persuade Sky of his views: 'I 
sent them lengthy fax messages explaining it to them in baby 
language,' he recalls. 

Sky's technical guru, Peter Smith, also backed the smart card, 
and Murdoch gave in under the pincer movement from Sugar and 
Smith. One day he picked up the phone and told Sugar, 'Don't 



296 



ALAN sugar: the amstrad story 

start barking like an old crow and shouting your head off, but 
basically you were right. We're going with the smart card.' 

Sugar naturally hoped to win the contract to make the smart- 
card descrambling system, but he was thwarted for two reasons. 
The first - and, Sky insists, most important - was because Peter 
Smith was not convinced that Amstrad's smart card was secure. 
The second was a matter of timing: Murdoch wanted the descram- 
bler to be available by April 1989. Sugar sat down with his 
engineers and decided the deadline was impossible. He told 
Murdoch that Amstrad could deliver the smart card by Septem- 
ber, whereupon Murdoch informed him that Thomson, the French 
electronics giant, had promised to deliver by June. Murdoch was 
deluged with more Sugar candour: 'When hair grows in the palm 
of your hand, then Thomson will deliver in June. Go to Thomson, 
those great French mega-stars, and you will not scramble until 
into 1990.' 

Thomson got the contract and failed to deliver in June 1989, 
since the work on the scrambler proved more difficult than either 
it or Sky had anticipated. As a result, Sky had to postpone the 
launch of its scrambled movie service until 1990. Moreover, the 
scrambling discussions held up work on the design and production 
of Amstrad's dishes in the autumn of 1988, a critical period in the 
run-up to Sky's launch. 

So too did a more technical but related issue. Sky needed to 
issue a specification of the socket which would allow the decoder 
to be plugged into the receiver. The difficulty was that this 
specification had to be suitable for equipment from the twenty or 
so companies which were thinking of manufacturing receiving 
equipment for Sky. The latter's starting point in writing the 
specification was Amstrad, since it would be the first and largest 
receiver manufacturer. But having determined what Amstrad 
wanted, Sky then had to ask the views of all the other potential 
manufacturers. 'It was over a month while they were sodding 
around making their mind up which way they were going to go,' 
Sugar recalls. 

Sky's planners had a bad moment on 10 December when the 
Ariane-4 rocket was sitting on its launch-pad in Kourou, French 



297 



DAVID THOMAS 

Guiana, with the Astra satellite on board. Already postponed once 
from November, the launch was aborted just 26 seconds from lift- 
off when a faulty valve was discovered. Twenty-four hours later, 
however, the rocket blasted off perfectly, carrying into the strato- 
sphere Murdoch's and Sugar's hopes. 

Preparations were feverish at Sky's new £15 million studios on 
the western outskirts of London as the February 1989 start-up 
date approached. Less than a week from launch, visitors found 
parts of the forecourt still a sea of mud. Meanwhile, Murdoch 
scaled back his ambitious forecasts for Britain's first satellite 
television service. In January 1989, Sky told potential advertisers 
that it was cutting predictions of the minimum number of homes 
it expected would be receiving Sky after a year from 2.5 million to 
1.15 million. One problem becoming obvious to friend and foe 
alike was that there would be virtually no receivers in the shops at 
launch date. 

When Sky began broadcasting, retailers like Dixons and Currys 
typically had only one dish in each store which they used for 
demonstration purposes. Murdoch conceded that most potential 
Sky customers would have to wait for more equipment to be made 
by Amstrad's Japanese sub-contractors - 'the night shift in Japan', 
as he put it- before they would be able to find a receiver. The delays 
imposed on Amstrad by Sky the previous autumn came home to 
roost and the build-up to full-volume production of receivers by 
Amstrad's contractors in the Far East did not occur until a month or 
so after Sky's launch. Sugar fumed at the suggestion that his 
company was responsible for the delay; Murdoch, for one, is in no 
doubt as to where the blame lies: 'It was my fault. That wasn't 
Amstrad's fault. That was my fault,' he says in retrospect. 

Sky's launch went without a hitch, but there was plenty of sniping 
in that portion of the media not owned by Murdoch about the 
quality of the programmes on offer. For some unfathomable reason, 
the show of country singer Dolly Parton was a particular target. 
Nevertheless, even the most hostile critic had to concede Sky's 
achievement in simultaneously launching four new television chan- 
nels - the first time this feat had been attempted in Britain. Sugar 
was in Japan on the day of the launch, but when he returned the 



298 



ALAN sugar: the amstrad story 

following weekend he switched on Sky at his home, one of the few 
in the country then sporting an Amstrad dish. 'I'm impressed,' he 
told a group of City analysts a few days later. 

Murdoch's newspapers lost no chance to puff Sky: the Sun's 
page 3 girl even appeared clutching a satellite dish and little else. 
This campaign, together with the general publicity surrounding 
the launch, unleashed a surge of demand for the dishes. Unsur- 
prisingly, greatest interest was found among the skilled working 
class, the C2s, the bedrock of Amstrad's traditional consumer 
following. 'People are driving us barmy wanting the dishes,' Sugar 
said, while Murdoch claimed that retailers had taken deposits 
from almost 100,000 people. 

But there were still not enough receivers. Amstrad considered 
flying in a batch from the Far East, thereby cutting out their long 
sea journey, but that would have been an expensive solution for a 
product on which its profit margins were already tight. Sugar 
decided that the retailers would have to pay to air-freight the 
receivers: 'One of us is going to have lower margins, either us or 
the retailer - and it won't be us,' he said. 

Sky managed another early publicity boost by screening the 
heavyweight title fight between Mike Tyson and Frank Bruno live 
at the end of February. But the lack of dishes in the shops 
prevented it from really cashing in on this coup. Daniel, Sugar's 
younger son, displayed the family's entrepreneurial streak by 
hiring a hall near the Sugars' Chigwell home. By charging the 
local populace £30 a head for a meal and a chance to view the 
fight on a large screen hooked up to Sky, the 18-year old made 
some useful pocket money. 

However, the initial surge in demand for the dishes proved to 
be short-lived and interest dropped sharply soon after Sky's 
launch. Sky also had to cope with the embarrassing 'melting dish' 
story. G.E.C. had made a blue plastic cover for its LNB, but one 
batch of the covers lacked ultra-violet inhibitors. Unfortunately, 
Britain was lounging in one of the hottest summers on record and 
the plastic covers disintegrated in the sun. Amstrad and G.E.C. 
moved quickly to remedy the defect. 

In April, the lack of consumer interest forced Sky to shed half 



299 



DAVID THOMAS 

its advertising staff'. By June the Financial Times' $ monthly satellite 
monitor was reporting that at most 110,000 homes had installed 
dishes - well under 1 per cent of all British households. Comedians 
added 'Sky jokes' to their repertoire: 'What's the difference 
between the Loch Ness monster and Sky Television?' ran one. 
'Some people claim to have seen the Loch Ness monster!' 

The Sky camp blamed its lack of progress on the hostility of the 
non-Murdoch media. It was also alarmed by a water-muddying 
advertising campaign launched by B.S.B. Sky's rival appeared to 
have successfully confused many people into deferring a decision 
about whether or not to try out this new satellite phenomenon. 

One chink of light for Sky appeared in May, when B.S.B. was 
forced to postpone its launch into 1990. To its embarrassment, 
B.S.B. was in the middle of a £20 million advertising campaign 
telling the British public about its plans to launch in September 
1989. But no sooner were the commercials placed than it became 
clear that if it went on air at this date it would have virtually no 
dishes for several months. Development work on the chips for 
B.S.B.'s dishes by I.T.T., the U.S. -based conglomerate, was way 
behind schedule. Two and a half years after B.S.B. was awarded 
its franchise, it was still not ready to transmit. Sky, which had 
launched on target less than a year after announcing its plans, 
could not resist the temptation to gloat. 'Sky: on air. B.S.B,: hot 
air,' it proclaimed in newspaper adverts. 

Sky would have the market to itself for at least another nine 
months, including the crucial Christmas period, so Murdoch 
decided to milk this opportunity for all it was worth. But by this 
time relations between Sky and the high street retailers had 
deteriorated. Sky was accusing the retailers of not doing enough to 
promote the dishes, while the retailers replied that the demand 
simply did not exist. Sky decided to break the impasse by supplying 
the dishes direct to customers: 'We went into the business of leasing 
the dishes ourselves. That was after we'd had big rows with Dixons 
and all of the multiple chains,' Murdoch explains. 

Sky offered customers a promotional deal, by which they would 
pay £4.49 a week for a package including the installation and 
maintenance of receivers and a subscription to Sky. Installation 



300 



ALAN sugar: the amstrad story 

companies were appointed around the country to handle the 
business. Sky also had to buy large numbers of dishes, so Murdoch 
asked Sugar to call in at his office at News International's 
Wapping headquarters. 

Sugar had to walk through the Sun news-room to reach Mur- 
doch's office. 'You are now entering Sun country' is emblazoned 
above the main entrance, while blown-up versions of some of the 
paper's more infamous front pages are plastered over the walls. 
Sugar went through another security door - a Wapping feature - 
past the hexagonal waiting area littered with copies of Murdoch's 
magazines and into his vast office. Neither Murdoch nor Sugar 
are slow decision makers. It took less than twenty minutes for 
Sugar to emerge with an order for about 500,000 dishes to be 
delivered over the following six months. 'We wanted to back him. 
We thought we owed it to him. He'd taken great risks. And we 
still trusted him to give us the most reliable delivery. Everyone 
else had been talk,' Murdoch notes. 

The trade-off for such a large order was a keen price: 'Well, we 
thought it was a good price. Like a good salesman, Alan tried to 
convince me that it would only just cover his costs. But I hope 
he's making something out of it,' Murdoch says benignly. 

Sky backed the rental offer with a £21 million advertising 
campaign. The renewed burst of publicity after it was announced 
in September also rekindled interest in straight sales of the dishes, 
an unexpected spin-off which helped smooth a few ruffled feathers 
among the retailers. By the middle of December take-up of the 
rental offer was running at 4,000 a day, all of them Amstrad 
dishes. 

As 1989 drew to a close, Sugar turned his mind to selling his 
dishes into Germany. But first he had to deal with an embarrassing 
side effect from Astra's German services for Sky's operations in 
Britain. 

Early in December, Astra began transmitting four new German 
channels which could be received on Amstrad dishes, but due to a 
technical oversight on Astra's part the Amstrad receivers did not 
pick up the German sound. As a result, owners of Amstrad dishes 
throughout Britain were suddenly in the puzzling position of being 



•301 



DAVID THOMAS 

able to view four new soundless channels. Every satellite television 
service centre in Britain was flooded with complaints and queries. 
Once their switchboards became clogged, the viewers began to 
bombard Sky with calls. 

Amstrad and Sky both tried to phone Astra in Luxembourg, but 
everyone at the satellite company seemed to be out celebrating the 
successful launch of their German channels. Sugar then called 
Peter Smith at Sky and they decided they would both send urgent 
faxes to Astra. 

Smith dictated a fax to his secretary who queried it: 'That's a 
little bit terse, isn't it? 1 she said. 'No, I think it's O.K., send it,' 
Smith replied. Sugar then sent his fax to Astra, with a copy to 
Smith; it was a message of such directness that it took the 
Australian's breath away, but the two-pronged attack worked: 
Astra fixed the problem within 24 hours. 

Amstrad's chances of selling large quantities of dishes on the 
Continent vary from country to country, depending on factors 
such as the broadcasting system and the extent of cable television 
in each one. But Germany, the Netherlands and Scandinavia 
should hold out some opportunities. In Britain so far, the satellite 
story has been mixed. 

Clearly sales have not matched the early expectations of either 
Amstrad or Sky. The weak market has allowed Amstrad's imme- 
diate customers - both the retailers and Sky - to drive tough 
bargains with Amstrad, putting pressure on the profit margins. 
Yet Amstrad still managed to sell 670,000 dishes in 1989. As both 
Amstrad and Sky never tire of stressing, the initial take-up of 
satellite television has been much more rapid than that of previous 
consumer electronics innovations, such as the video recorder. 
Moreover, Amstrad had mopped up most of the business that was 
available: according to Sky, it accounted for 80 per cent of the 
dishes installed in homes by the end of 1989. 

At the end of Sky's first year, Sugar took a ruthlessly realistic 
view of the future: 

What's happening now is you're starting to see the Japanese 
and other people coming in with reasonably priced equipment. 



302 



ALAN sugar: the amstrad story 

There will be a lot more choice of satellite equipment soon. 
Then Amstrad's market share dominance will start to go down, 
margins will start to erode and we'll say: thank you very much, 
we're on our BMX bike and we're off, we're out of this market, 
just like we were out of CB [Citizen Band radio] and 14" tele- 
visions. In two to three years time, the old moss will start growing 
and Mr Murdoch will not remember me. He won't even know 
who I am, because there will be 500 different suppliers of satellite 
receivers out there. But that's life. That's why we had to be 
realistic. We had to get in and make some money. We have to 
realize that we haven't got a monopoly. Once we've shown the 
way to others, everyone will jump on the band-waggon. So don't 
be upset about it and don't expect any sentiment. 

This may in fact turn out to be unduly pessimistic. True, many 
other players have now launched dishes; but overall sales are likely 
to expand steadily, assuming that satellite television follows the 
typical growth path in consumer electronics. Moreover, at least 
some people at Sky believe that Amstrad will remain in a strong 
position: 'I don't see their market share falling rapidly. I think 
they'll maintain a leading role/ says Peter Smith. 

The satellite television market will evolve as it becomes more 
sophisticated, throwing up opportunities for Amstrad to produce 
new products. An early example is the combined receiver-decoder 
on which the company was busy working in the first few months 
of 1990. 

Neither does Rupert Murdoch seem likely to forget Alan Sugar 
as the latter predicts. Murdoch believes that Amstrad's ability to 
produce large numbers of satellite dishes cheaply, when everyone 
else was dithering, fully justified his early generous assessment of 
Sugar: 'He's very entrepreneurial, a tremendous worker. In nego- 
tiations, he's a master of detail. I found he came to the point, to 
the bottom line very quickly. He's been very straight with me - 
totally. He's kept his word on everything.' 

Murdoch was not ruffled by the fact that Amstrad was going 
through a bad patch in other areas of its business immediately 
before and after Sky's launch: 'It didn't affect his behaviour 



303 



DAVID THOMAS 

towards me at all. Why should it? I had a bad year myself/ says 
Murdoch. 

Indeed, while Murdoch believes that Sugar's share of the dish 
market is bound to decline, he envisages working with Amstrad on 
future projects: 'No one knows where electronic communications 
are leading to. I'm sure there will be lots of other electronic 
devices, whether they're all built together into one entertainment 
centre, whether there are printers out of the side of high-definition 
television. Who knows? But I'd be surprised if there aren't other 
things we do with him.' 

Amstrad participated in one of the most expensive entertain- 
ment and consumer electronics projects ever undertaken in 
Europe without exposing itself to great risk. Sugar ensured that 
others bore the expense of launching the satellite broadcasting 
revolution, ducking out of B.S.B. as soon as it became clear that 
continuing involvement would demand considerable cash with 
little immediate payback. In the case of Sky, the television 
company itself shouldered most of the risk. Sky was losing up 
to £3.5 million a week in 1989, cutting swathes through the 
profits of News Corporation, Murdoch's Australian holding com- 
pany. By contrast Amstrad has made money out of satellite 
broadcasting, even if not on the scale that Sugar had initially 
hoped. 

Meanwhile Sugar watched B.S.B.'s progress with growing 
incredulity, thankful that he was not still a shareholder. Late in 
1989, having already secured £423.6 million from its backers, 
B.S.B. proposed to raise a further £900 million from a mixture of 
new shares and loans. Sugar could not understand why the City, 
which was so alert to every blip in Amstrad's trading performance, 
could blithely stand by while B.S.B. raised such vast sums for a 
venture which would take years to make a profit. 

The City's treatment of B.S.B. exasperated Sugar. By the end of 
1989 it was driving him to the brink of becoming a patron of the 
arts, as he explains: 

If the City allows them to pour all that money into that thing, 
then how they've got the audacity and cheek to complain about 



304 



ALAN sugar: the amstrad story 

Amstrad, I don't know. If they pour all that money into B.S.B,, 
that will be the last straw as far as I'm concerned. I'm going to 
commission a sculptor to produce a great big cement statue of a 
pair of fingers sticking in the air. And I'm going to hire a 
contractor to erect the statue in some street right in the middle 
of the City. 



305 



• 16- 
The Year of Disaster 



The thirty City analysts trapped in Liverpool Street station one 
day in mid-February 1989 were becoming impatient. They had 
fought through the tunnels and trenches which criss-crossed 
London's main East-line terminus; they had defied the lack of 
signposts to find platform 16; but there, in the filthy cavern which 
passed for a platform, they had to admit defeat. 'No idea,' was the 
porters' shrugged reply to queries about the whereabouts of a train 
to Brentwood. 

The well-bred complaints mounted as the minutes ticked away. 
'Why can't Sugar come to London like everyone else?' the analysts 
muttered, nervously fingering their portable phones. 

For a dreadful moment, it looked as though the City was going 
to miss its taste of blood. That morning Amstrad had declared the 
first fall in profits in its history as a public company. Pre-tax 
profits for the six months to the end of 1988 - the most important 
half of Amstrad's trading year - were down 16 per cent to £75.3 
million on sales almost unchanged at £348.8 million. The analysts 
who followed Amstrad for their stockbroking firms would have 
walked to its Brentwood headquarters to hear Alan Sugar's 
explanation. 

Such a drastic remedy proved unnecessary, because British Rail 
eventually managed to find a train. After the thirty-minute jour- 
ney, the analysts swept into Amstrad's headquarters, expecting to 
discover the company flat on its back. They actually found a 



306- 



ALAN sugar: the amstrad story 

chastened but chirpy Alan Sugar; Amstrad's chairman was 
relaxed, funny and remarkably honest when presenting the com- 
pany's results. 

'We can't afford a launch any more,' Sugar replied to a query 
about plans for new product launches. Unprompted, he told the 
gathering that he had wanted to call the financial year 1988-89 
'the year of disaster', but his advisers had vetoed the idea. Just in 
case anyone had missed the point, Sugar repeated it: this is 
Amstrad's year of disaster. 

The directors had dissected the factors behind the decline in 
profits. 'We've been critical of ourselves. We've tried to understand 
what went wrong, so we can decide what to do,' Malcolm Miller, 
Amstrad's marketing director, told the analysts. 

The essence of the problem was that Amstrad had not been able 
to make enough of its products in the second half of 1988. This left 
a pool of demand which the company could not satisfy. Amstrad 
had worked out the precise gap between the machines it had at its 
disposal and those it had expected to make, figures which Miller 
reeled off to the analysts: 90,000 business computers worth £57 
million in sales - lost; 45,000 home computers worth £8 million - 
lost; 22,000 printers worth £3 million - lost; 190,000 audio units 
worth £22 million - lost; and £24 million worth of video recorders 
- lost. 

In total, Amstrad had lost £114 million in sales because it did 
not have the machines to meet demand. If those sales had been 
booked, then half-year profits would have been over £100 million, 
comfortably sufficient to maintain the company's upwards 
momentum. 

But this explanation, thorough as it was, immediately prompted 
other questions: Why had Amstrad lost these products? What had 
gone wrong with Amstrad's famed manufacturing flexibility? 
Sugar identified two different kinds of reasons when he pointed to 
'a chain of events, some outside our control and others the results 
of our own mistakes'. 

Consider, first, the events which were largely outside Amstrad's 
control. The most important external pressure with which Sugar 



307- 



DAVID THOMAS 

grappled in 1988 was one of the periodic bouts of shortage in the 
supply of memory chips. 

Memory chips - and especially devices known as DRAMs 
(dynamic random access memories) - are one of the most import- 
ant components in a computer. Needed in increasingly large 
volumes to store the data processed by a computer, memory chips 
are the work-horses of the semiconductor family. Unfortunately, 
their supply tends to swing unpredictably from feast to famine and 
back to feast again with almost biblical savagery. 

The memory chip and computer cycles feed off one another in a 
perpetual loop. Exceptional demand for computers results in 
strong orders for memory chips; the order books of DRAM 
manufacturers fill up, delivery times lengthen and the price of the 
chips rises; escalating chip costs force computer manufacturers to 
put up their prices, which depresses the demand for computers 
just as chip manufacturers are churning out DRAMs at maximum 
volume; the market becomes flooded with memory chips, leading 
to a cut in DRAM prices; this in turn feeds through into cheaper 
computers and a resurgence of exceptional demand for computers. 
And so on, potentially ad infinitum. 

Towards the end of 1987, the DRAM cycle was exacerbated by 
political tension between Japan and the United States. In the first 
half of the 1980s, Japanese companies had identified control of the 
memory chip market as a key to world leadership in electronics. 
Huge Japanese investments in memory chip capacity allowed 
them to offer DRAMs to American and European computer 
manufacturers at rock-bottom prices, sweeping aside the compe- 
tition in the process. 

The United States Government watched the Japanese assault 
with growing alarm. Semiconductors are the modern equivalent of 
steel: they are fundamental to a country's ability to wage successful 
warfare. No military computer can run without memory chips. 
Goaded by cries from American semiconductor makers that their 
Japanese competitors were dumping DRAMs at unfair prices, the 
United States pressured Japan into signing a semiconductor trade 
pact in 1986. 

The pact forced Japanese chip manufacturers to increase 



308 



ALAN sugar: the amstrad story 

DRAM prices. They also throttled back their production, cutting 
the supply of memory chips from Japan. Unfortunately, their 
drive to dominate the world memory markets had been so 
successful that few chip makers were left outside Japan to fill the 
resulting supply gap. The large U.S. semiconductor companies 
had dropped out of DRAMs one by one: by the end of 1987 there 
were just two significant American companies selling DRAMs on 
the open market, Texas Instruments and Micron Technology. 
Japan had already captured more than 70 per cent of world 
DRAM supply. 

Superimposed on the normal memory chip cycle, the U.S.- 
Japan semiconductor pact had a dramatic effect. DRAM supplies 
dried up and prices rose. Alan Sugar told the analysts gathered at 
Amstrad headquarters that he smelt the first whiff of an impending 
DRAM crisis in December 1987: 'But I hoped it was going to go 
away. I'd heard that story before. You get this waffle all day long 
in the semiconductor industry. I didn't believe it.' 

This time, however, it was not waffle. Amstrad was used to 
paying about $2.50 for a 256k DRAM (a chip capable of storing 
256,000 pieces of information). By April 1988, the company was 
paying $5 as a matter of course. 'In some cases, we were so 
desperate, we paid $8 or $10 - blackmail prices,' said Sugar. 

Between June and October 1988, he spent most of his time on 
the DRAM problem. He searched the globe for alternative sup- 
plies, which meant scouring the Far East and the United States, 
since Europe was an also-ran in the memory chip race. A type of 
memory chip from the 256k family, known as a 64k by 4 chip, was 
vital to the graphics capabilities of Amstrad's computers; yet this 
chip was in particularly short supply: 'That was the bloodline,' 
Sugar told the analysts. 

Spiralling DRAM prices slashed into Amstrad's profit margins. 
Each PC 15 12 personal computer contains 18 256k DRAMs, while 
each PC 1640 contains 30 of the chips. DRAMs at $5 each would 
gouge $75 from Amstrad's profit on a PC 1640, while a $10 DRAM 
did unthinkable things to the company's bottom line. Sugar 
reacted by stopping production of the cheaper single disk drive 
versions of the PC 1512 and PC 1640 and cutting back on output of 



309 



DAVID THOMAS 



home computers, concentrating the company's efforts on the 
higher priced computers in its ranges. This - together with some 
price rises on Amstrad's products - helped to stem the erosion of 
its margins. But overall sales and hence profits were inevitably hit. 
Sugar reckoned that half the £57 million of lost sales in business 
computers stemmed from the DRAM shortage. 

Of course, Amstrad was not alone in having to face the DRAM 
crisis. Two weeks before the analysts trekked out to Brentwood, 
Apple Computer warned that its profits would fall by 29 per cent 
in the first quarter of 1989 because it had been forced to buy 
memory chips on the spot market at three times the normal prices. 
On the very day that Amstrad released its results, star German 
computer company Nixdorf disclosed that it would be slashing its 
dividend by 60 per cent to compensate for higher chip costs. 

But the fact that others were in the same boat was scant comfort 
for Sugar who felt the need for a stable solution as DRAM 
shortages worsened in 1988. All the signs pointed to a prolonged 
crisis. In July 1988, Dataquest, the respected U.S. -based electron- 
ics research organization, estimated that only 20 million 256k 
DRAMs would be available in Europe in 1988 - 5 million pieces 
short of the likely demand for these components. Even companies 
buying in large volumes were having to pay up to $15 for a 256k 
DRAM on the spot market. By September the memory chip 
famine was compounding delays of the launch of Amstrad's newest 
computer range, the PC2000 series. 

Sugar decided to reduce Amstrad's vulnerability to this see- 
sawing market by seeking a stable source of memory chips. In 
October 1988, Amstrad announced that it had paid £45 million 
for a 9 per cent stake in Micron Technology, then one of the two 
remaining U.S. producers of memory chips for the open market. 
The deal gave Amstrad the right to buy up to 9 per cent of the 
Idaho-based company's output of chips, although it was free to 
source from elsewhere if it could find cheaper supplies. Amstrad 
had also signed a one-year contract for DRAM supplies with 
Samsung, the South Korean electronics giant. 'There is no good 
done by sitting around, whimpering and hoping that the problem 



310 



ALAN sugar: the amstrad story 

will come right. We had to be decisive and do something,' Sugar 
said at the time. 

Amstrad's stake in Micron was hailed as 'a smart investment' by 
a U.S. investment analyst in the Wall Street Journal, but Sugar was 
later to regret the decision. Soon after the Amstrad-Micron link was 
announced, memory chip supply improved. Factories in the Far 
East began to churn out the new, more powerful one megabit DRAM 
- capable of storing one million pieces of information. By the end of 
1989, the world was flooded with one-megabit DRAMs and memory 
prices collapsed. Whereas in 1988 one megabit DRAMs had cost as 
much as $30 each, by August 1989 they were selling in the United 
States for about $14.50, while by the start of 1990 they were trading 
for as little as $6.50. Micron's stock price reflected the market for 
one of its main products. By October 1989, the value of Amstrad's 
stake in Micron had fallen by about 30 per cent after allowing for 
currency movements. 

Buying the stake in Micron was not 'the greatest deal I've ever 
done', Sugar acknowledged at the analysts' meeting in February 
1989. Yet it did serve to extricate Amstrad from the nightmare of 
the DRAM shortage. The episode also taught Sugar that Amstrad 
had become too big to risk similar exposure to the volatile 
semiconductor market in future: by the end of 1989, he had deals 
in place with five chip manufacturers - Micron, Samsung, Sie- 
mens, Texas Instruments and Toshiba - which secured Amstrad's 
future supply of memory chips. 

The DRAM shortage fed through into fewer sales of Amstrad's 
computers, which in turn cut demand for the company's printers. 
Its other product lines were affected by their own problems, again 
largely outside Amstrad's control. When labour shortages in the 
feverish Taiwanese economy disrupted the flow of audio products 
from Amstrad's sub-contractor there, Avnet International, Sugar 
summoned Avnet's top man from Taiwan to see him in Brentwood 
and extracted promises that the volumes would improve. But they 
did not, and Amstrad did not have the number of audio units it 
wanted for the 1988 Christmas season. The experience reinforced 
Sugar's growing determination to cut Amstrad's dependence on 
Taiwan, which was rapidly becoming a high-cost production 



•311 



DAVID THOMAS 

centre. In Hong Kong, Stan Randall negotiated a supply of low 
technology audio parts from factories in Communist China, while 
some of Amstrad's Taiwanese sub-contractors moved their pro- 
duction facilities to Malaysia with its encouragement. 

Meanwhile, the Funai-Amstrad joint venture to make video 
recorders in Shoeburyness was proving 'a disaster', in the words 
of an Amstrad insider. One view within the company was that the 
Funai side - which was in charge of running the plant — put less 
effort into the operation than if it had been a wholly owned 
subsidiary. Whether or not this was true, the fact was that the 
factory was slow to build up production. Amstrad hoped to be 
sourcing all its video recorders from Shoeburyness by the second 
half of 1988, but as the year progressed it became increasingly 
apparent that it would not have the volume of video recorders it 
needed for the Christmas season. 

Although Amstrad piled the pressure on Funai, the result was 
less than happy, as Jim Rice - the Amstrad director with day-to- 
day oversight of the joint venture - describes: 

The problem was, as always with Amstrad, it was never quick 
enough. We decided we wanted more product out of the factory, 
but the Japanese in their normal Japanese way were doing 
things properly. So along we came and said, 'Sorry, this is not 
good enough, we need more, and if you can't produce more we 
will put our own people in to do so.' That's what we did and we 
began to produce more and more. But then we began to suffer 
from quality problems, labour problems, all sorts of things - 
because we were trying to rush it. 

The episode confirmed Sugar's distaste for direct involvement in 
manufacturing. Once the 1988 Christmas season was over, 
Amstrad's chairman sat down and said to his team, Tve had 
enough of Funai-Amstrad. It's chaos. We don't want to be 
involved with it any longer. Tell Funai they can take over the 
entire plant.' Amstrad set about selling its 49 per cent share in the 
joint venture to Funai, although it still planned to take large 
volumes of video recorders from them. 



312 



ALAN sugar: the amstrad story 

The DRAM crisis, labour shortages in Taiwan and Funai's 
delays in building up production - these were factors largely 
outside Amstrad's control. But two further problems underlay the 
company's profit dip, for which only Amstrad was to blame. 

The first was poor management of the ending of the West 
German distribution agreement with Schneider. When Amstrad 
severed its relations with Schneider, it left the German company 
with large volumes of Amstrad computers badged with the Schnei- 
der name, which Schneider proceeded to sell in Germany at 
discounted prices just as Amstrad was trying to build up its 
subsidiary there. The Schneider stock undercut Amstrad's own 
computers, disrupting Amstrad's re-launch in West Germany in 
1988. Sugar accepted full responsibility for this uncharacteristic 
mistake: 'We accept in hindsight that this situation was self- 
inflicted. We clearly should have taken control of their inventory.' 

The second problem - lengthy delays in introducing its new 
flagship PC2000 computer range - was even more serious. 
Amstrad had originally hoped to launch the line in February 1988, 
but Sugar was not able to unveil it until September. Moreover, he 
then had to disclose that the two most expensive computers in the 
range - the PC2286 and PC2386 - would not be available in the 
U.K. until December 1988. Sugar blamed these delays on the 
memory chip shortage. 

The December deadline came and went with still no sign of the 
PC2286 and PC2386. By then it was apparent that problems 
within Amstrad were holding up the computers - problems which 
accounted for the other half of the £57 million of lost business 
computer revenue. At the mid-February meeting, Malcolm Miller 
told the analysts that shipping of the PC2386 would begin on that 
very day, but the PC2286 would not be shipped until March. As a 
result, the new series would not be fully available in the U.K. until 
May or June. Faults had been discovered in one of the chips which 
had been specially designed for the PC2000 range. Repeating its 
standard trick, Amstrad had tried to cram a large number of 
functions on to a couple of chips, but the PC2000 range contained 
more sophisticated technology than anything previously attempted 
by the company. In retrospect, Sugar concluded that it would 



313 



DAVID THOMAS 

have been better to spread the functions around more chips, even 
if this had added to the cost of the computers. 'It was our own 
fault. We were being very, very adventurous. We took on more 
than we could chew/ he admitted to the analysts. 

However, Sugar's performance in front of the analysts demon- 
strated a willingness to recognize mistakes and to tackle them - a 
quality not always abundantly present in business stars who have 
faltered. He answered with characteristic frankness the half-dozen 
or so long-time Amstrad followers whose well-informed questions 
punctuated the meeting. Yet they were also alarmed by the 
message that they should not expect any improvement during the 
second half of financial year 1988-89. Immediately after the 
meeting the analysts picked up their mobile phones to telephone 
in downgrades of their profit forecasts. Amstrad's share price 
plunged 22*/2p - a 12 per cent fall during the day - before coming 
to rest at 158V2p, wiping £57 million off the value of Sugar's own 
holding in the company. 

The analysts relaxed on the train back to London, preparing 
themselves for another encounter with Liverpool Street station. 
One of the youngest of the group, who had sat silent throughout 
the meeting with Amstrad, turned to his neighbour and asked 
sotto voce, 'What exactly is a microchip? Where would I buy one 
if I wanted one?' 



Worse was to come. When he addressed the City analysts in 
February 1989, Sugar had hoped that the year of disaster was 
drawing to a close. But some problems had not even broken. In 
the next nine months, he was forced by the gathering storm clouds 
to confront the most serious crisis in his business career. The year 
of disaster turned into 18 months. 

The discovery of technical hitches in the PC2000 series was the 
most obvious hammer blow. These computers represented a 
renewed assault on the corporate computer market. The top end 
of the range - the PC2286 and PC2386 - used the Intel 80286 and 
80386 microprocessors, more powerful semiconductors than 



314 



ALAN sugar: the amstrad story 

Amstrad had designed around before; yet, as one would expect 
from the company, prices were pitched below the competition. 

Even so, the price tickets on the range showed just how far 
Amstrad had moved beyond the truck-driver-and-his-wife. The 
PC2086 cost £599 to £1,349; the PC2286 cost £999 to £1,699; and 
the PC2386 cost £2,649 to £2,999. By Amstrad's previous stan- 
dards, the top end of the range consisted of big ticket items. 

As powerful personal computers aimed squarely at the business 
market, the PC2000 range had an abundance of features for 
computer reviewers to crawl over. Inevitably, the reviews were 
more mixed than in the case of Amstrad's earlier and simpler 
machines; a reviewer tended to like some features, but not others. 
Nonetheless, the balance of comment in the main computer 
magazines was favourable when the series was announced in 
September 1988. 

Personal Computer World, for example, described the range as 
'competitive if not breathtaking', adding that 'no corners have 
been cut in performance to get down to a price'. Reviewing the 
entry level PC2086 machine, PC User commented, 'Amstrads are 
no longer the cheapest machines on the block - and competition 
from quality clones has never been tougher.' Yet still it concluded, 
'With its distinctive, space age look, good graphics and competitive 
price, the PC2086 deserves to do well - both among small 
businesses and among corporates looking for a low-cost intelligent 
work station.' 

In short, the message from the technical reviews was that 
Amstrad had every chance of success with its PC2000 series. These 
cautiously optimistic projections were thrown into turmoil - first 
by the delayed launch of the top end of the range and then by 
persistent rumours of 15-20 per cent fault rates among the PC2286 
and PC2386 once they saw the light of day. Unlike the overheating 
rumours which had previously beset the PC 1512, Amstrad 
accepted that something was wrong: the problem was in the circuit 
which controlled the hard disk drive. 'Unfortunately it turned out 
to be not something that could be fixed with a bit of external 
tweaking - there was something seriously wrong with the core 



315 



DAVID THOMAS 

logic inside the disk controller chip that we had designed/ Sugar 
said. 

In April 1989, Sugar suspended shipments of the PC2286 and 
PC2386. In a bid to sort out the problem, he called in Western 
Digital, the U.S. company which is the world's largest supplier of 
hard disk drives. Western Digital replaced the faulty chip with a 
standard one of its own; not prepared to take any chances this 
time, Sugar asked P. & P. - a leading computer distributor - to 
evaluate its work. 

Sugar disclosed the problem and the actions taken to correct it 
in late July 1989. He ordered the recall of the entire 7,000 hard 
drive models in stock and signed up Dictaphone, a maintenance 
company, to replace users' faulty machines. All PC2286 and 
PC2386 users were given a free one-year on-site maintenance 
contract with Dictaphone. 'A dramatic gesture is now needed to 
overcome any suggestion that this company is failing in its duty to 
supply reliable products,' Sugar stated at the time. 

Amstrad won many plaudits for the speed and honesty with 
which it handled the fault. Sugar estimated the cost of modifying 
each recalled machine at £75, but added that tumbling memory 
chip prices more than compensated for the extra cost. He displayed 
his most confident face after announcing the recall: Tm now 
convinced the machines work perfectly, and we're geared up for 
the bomb-blast Amstrad effect, creating a mass market as we have 
done so successfully before.' 

Yet worrying questions were raised both within and outside the 
company by this episode. Why was Amstrad so late in launching 
the PC 2000 series in the first place? And why did it fail to discover 
the hard disk fault until after it had launched this flagship range? 
The simple answer is that Amstrad's small design team was 
overwhelmed by the magnitude of the task. 

Once more MEJ Electronics had been asked to design the basic 
circuitry for the PC2000 series. Adopting the approach which had 
worked so well with previous Amstrad computer designs, Mej set 
about embodying many of the computers' features into the gate 
arrays. One example among many was the cache (a form of 
sophisticated memory) controller for the PC2386, which Amstrad 



316 



ALAN sugar: the amstrad story 

could have bought off the shelf from Intel, the U.S. -based electron- 
ics company; but that would have been expensive, so Mej had to 
design a cache controller from scratch which could be embodied 
in a gate array. Looking back, Mej believes it was wrong not to 
think through more clearly the work entailed by the large number 
of innovations which were being built into the design: 'We threw 
a number of ideas into the specification which we thought were 
good ideas without looking hard at the implications for the 
development work. It was a mistake for us to put them in without 
sitting down and investigating how much additional effort they 
were going to take - and asking Amstrad whether it was worth 
putting in such a feature or better to stick to the original timescale.' 

Mej's company had expanded to about half-a-dozen people by 
mid- 1987, when it began to work on the PC2000, but this was still 
a tiny team to tackle such a large project. Effectively they were 
designing two distinct computers, since the PC2286 and PC2386 
were based on two different, albeit related microprocessors - the 
Intel 80286 and 80386. The task was an order of magnitude more 
complex than the designers' previous work for Amstrad. Mej soon 
found himself designing a 10,000-gate array; by contrast, the gate 
array in Amstrad's first computer, the CPC464, had only 1,159 
gates. Knee-deep in design work, Mej was unable to keep a 
sufficiently close eye on the overall project: 'What I should have 
been doing was overseeing the project and spotting the fact that 
we were putting additional stuff into the specification. I should 
then have asked Amstrad whether we should do that and extend 
the timescales and development costs, or not, as they preferred. 
It's easy to say with hindsight.' 

Mej's team fell badly behind schedule, but still managed to 
produce innovative, cost-effective designs which worked. MEJ 
Electronics was not to blame for the fault in the disk drive 
controller which forced Sugar to recall some of the range, since 
Amstrad had given the job of designing the controller to another 
company. 

For some time Amstrad had felt that it was unwise for a 
company of its size to be entirely dependent for its basic circuitry 



317 



DAVID THOMAS 

designs on Mej's small outfit. This was one reason for establishing 
its own research and development team under Richard Altwasser, 
the former Sinclair designer. But Amstrad decided to go one further 
with the PC2000 project and parcel the work out to more than one 
group of hardware design sub-contractors. Mej recognized the logic 
of Amstrad's decision. Indeed, this was shown to be right in principle 
by the fact that MEJ Electronics struggled with the remaining work- 
load. There was only one snag: Amstrad was unfortunate in its 
choice of designers for the hard disk controller. 

It was obvious to both Roland Perry, Amstrad's technical 
manager, and Mej that the hard disk design was inadequate when 
it was delivered in the summer of 1988. But by then Perry and 
Mej were engaged in the frantic effort to remove the bugs from the 
work of Mej's team. In retrospect, Perry believes that radical 
decisions at that stage might still have avoided the later need for 
the recall: 'We should have ditched the disk controller round 
about September 1988 and not tried to debug it. We should have 
simply recognized that it was in too ropey a state too near to 
production. We could have bought in a disk controller at that 
stage.' 

MEJ Electronics had shown itself to be a first-rate electronics 
design team, but by the end of the 1980s it was too small to bear 
the full strain of Amstrad's demands. Roland Perry had only half- 
a-dozen engineers working for him in his role as Amstrad's project 
manager, again probably too few to oversee the development of 
complex products like the PC2000 range. In addition, by the end 
of 1989 Amstrad had another 15-20 engineers in its own research 
and development department under Richard Altwasser - insuf- 
ficient to cope comfortably with the stream of products that 
Amstrad was designing in-house such as its home computers, a 
fax machine, a portable personal computer, modems and com- 
bined television-video products. Some outsiders said that the strain 
on the in-house team showed through in the poor results of some 
of the projects under its wing. One example was the PC200 games 
computer, a flop launched in September 1988 which computer 
journalist Guy Kewney damned in PC Dealer as 'Sugar's Christmas 
turkey'. 



318 



ALAN sugar: the amstrad story 

The most damaging aspect of the botched PC2000 launch was 
its impact on Amstrad's standing in the corporate market. The 
company continued to insist that many purchasers of its computers 
were 'corporate' in the sense that they were VAT-registered - 
small businesses and independent professionals. This was 
undoubtedly true. Many small businesses and other organizations 
which had been introduced to computing through Amstrad's word 
processors or earlier personal computer ranges were ready to 
upgrade to more powerful machines; they were still attracted by 
the traditional Amstrad value-for-money virtues embodied in the 
PC2000 series. 

Yet there is no doubt that Amstrad intended the PC2000 series 
to straddle both the small business and large corporate markets. 
Many large companies used Amstrad's previous top end computer, 
the PC 1640, at least somewhere in their operations, and Amstrad 
wanted to build on that base. Peter Horrell - the Amstrad account 
director at advertising agency Delaney, Fletcher, Delaney - 
directed the PC2000 adverts more squarely at computer pro- 
fessionals: 'The target audience was much more the data process- 
ing manager, the chief executive, the senior manager generally - 
people who would influence purchasing decisions on personal 
computers in larger companies.' 

Amstrad believed that it could use its successful mass market 
techniques in selling to large companies. Malcolm Miller, com- 
pany marketing director, explains: 

We don't ever want to have the sort of prestige that I.B.M. has 
and trade simply on that. We're not trying to create this aura 
round us which others further up the market have. What we 
want is a good solid image, with excellent features at an 
attractive price - product you feel secure and familiar with and 
you can buy from lots of outlets. We are trying to be innovative 
with designs, techniques and with price; and I think that will 
always be our philosophy - to bring the price of say the 386 
down, add more features, advertise it more aggressively, rather 
than cloud the product in an exclusive image and put it on a 
pedestal. 



319 



DAVID THOMAS 

Advertising executive Greg Delaney set out to reflect this strategy 
in the PC2000 campaign: 

Part of what Amstrad has always done is to educate people as 
to the advantages of new technology and what it can do for 
them. As Amstrad sells to corporate buyers and data processing 
managers, they will get the back-up material and the technical 
information they need. But our advertising will still try to reduce 
the impenetrable to the comprehensible. We're in the business 
of explaining benefits. For example: faster processors - what do 
they mean? As a philosophy, we have not stuffed our trade press 
ads full of impenetrable jargon. We may have included more 
detail, but we haven't allowed it to overwhelm the ad. You see 
lots of ads in the trade press that are written by boffins. Let's 
hope that Amstrad never goes down that route, because it would 
be a disaster. You would land up with boffin talking to boffin. 
But Amstrad is a mass market company. 

Unfortunately, the problems which beset the PC2000 were likely 
to be particularly offputting to big corporate buyers. These 
cautious souls, who spend large amounts of other people's money, 
are the customers most likely to be scared by the PC2000's 
technical glitches and least likely to focus primarily on price. 

As Alan Sugar never tired of pointing out, it is true that 
Amstrad was not alone in suffering technical problems. Inordinate 
delays in launching new machines and the discovery of faults 
necessitating recalls are endemic in the computer world. Tales 
emerge from within ultra-secretive I.B.M. of printers under devel- 
opment for seven years, while the PCjr - I.B.M.'s only serious 
attempt to capture the low end of the personal computer market - 
was a flop of the first order. Launched in November 1983, 350,000 
PCjrs were still in stock when I.B.M. stopped making the machine 
seven months later. 

The difference is, of course, that it would take an earthquake to 
shift I.B.M. from the corporate market, whereas Amstrad was still 
fighting to establish credibility there. The problems with the 
PC2000 series threatened to destroy the image which Amstrad had 



320 



ALAN sugar: the amstrad story 

developed thanks to its previous computer triumphs - the image 
of a producer of machines which were both cost-effective and 
reliable. People began to remember that by origin Amstrad was a 
low end audio company. 

Some computer commentators resurrected an old chestnut - 
whether Amstrad's whole approach was unsuited to the big corpor- 
ate market. One of the biggest cliches in the modern computer 
industry is that it is smart to be a systems seller, not a box shifter. 
Most money is made not by simply selling standalone computers to 
corporate customers, but also by setting up computer networks and 
supplying the software, training and so on that the networks require. 
Amstrad was not by inclination an obvious player in the systems 
game. 'Amstrad is just not the sort of company that can sell systems 
to corporate users,' said John Lettice, editor of computer magazine 
Microscope, shortly after the PC 2000 recall. 

This scepticism about Amstrad's chances in the corporate 
market did not seem to be shared by some of its main rivals. 
I.B.M. was not alone in cutting the prices of its nearest equivalent 
machines after the launch of the PC2000 series. Yet the flow of 
sceptical comment on Amstrad's ambitions in the corporate 
market continued. This was particularly badly timed for Amstrad, 
because it came after a period of turbulence in its relationships 
with other important groups of customers. Many observers 
believed that its deteriorating customer relations were contributing 
to Amstrad's problems. 

Throughout 1988 complaints had mounted from some computer 
dealers about the back-up service available from Amstrad. Some 
alleged that it was next to impossible to get spare parts in good 
time from the company's service organization, which was then 
based in Amstrad's Shoeburyness warehousing and factory com- 
plex. Looking back Jim Rice, the responsible Amstrad director, 
accepts the thrust of these criticisms: 'We had a disaster through- 
out much of 1988 with spare parts. There wasn't enough room 
given over to it, so there was little spare stock and no systems to 
make sure we never ran out. We ended up not being able to supply 
spare parts to our dealer base and getting a really bad reputation 
because of it.' 



321 



DAVID THOMAS 

Simultaneously, relationships with some of Amstrad's biggest 
high street customers were becoming fraught. Alan Sugar has 
always had a view of retailers which is vigorous even by his 
standards. He sees them simply as conduits passing on Amstrad's 
goods to the end user: 

Our attitude is that we are doing the retailer a favour, and if 
one won't buy, we'll find another who will. That means our 
products must, by definition, be winners and that's how we go 
about designing them. Then we take a view on our markets and 
decide how many we can sell. . . . We'll go to a retailer and tell 
him here it is, it'll be out on such and such a date, there'll only 
be this many available at the beginning, and it's up to you. 

This was said by Sugar in late 1985, at the height of his first word- 
processor triumph, but it was a view which he changed little over 
the years. A salesman to his fingertips, he understood the psycho- 
logical importance of what happens when a customer walks into a 
shop. His insistence on launching products in ranges flowed from 
this understanding: 

You get people hooked on computers by leading in with a £399 
price. They walk into a store and they ask for the £399 Amstrad 
computer, and then they see another next to it. So they enquire 
from the salesman: 

Customer: 'What's that one?' 

Salesman: 'Ah, that's the one with a hard disk drive and a 
colour monitor.' 

Customer: 'Oh. What does that mean?' 

Salesman: 'Well, a hard disk drive means that it holds more 
storage. You know, it's got 20 megabytes. You don't have to 
keep putting floppy disks in it. A colour screen means you can 
have bar charts on it. Bla-de-bla-de-bla. 

Customer: 'Oh. How much is that, then?' 

Salesman: 'Well, that's a grand.' 



322 



ALAN sugar: the amstrad story 

He's probably a business customer buying it anyway. For 
someone in a small business, the difference between £400 and 
£1,000 is nothing. While everyone thought we were selling a 
cheap £399 clone, the reality was we sold more high ticket 
products. That was our marketing ploy. It happened also in the 
days of our audio equipment. We used to have a £199 Tower 
Unit with a single cassette deck, and we would only sell it to 
dealers who would stock both our models - the one with the 
double deck as well. The idea was to put them side by side, the 
£199 next to the £260 one. It's absolutely normal for a member 
of the public to be lured to a shop by Amstrad's advertising of a 
£199 lead-in price point and to say, T want that Amstrad. Oh, 
what's this next to it?' The salesman Dixonizes them. He jumps 
on them and says, 'Well, that one is the all-singing, all-dancing, 
more powerful, double cassette, bla-de-bla.' Nine times out of 
ten the customer will pay it off on credit, which works out at £1 
a month more for the better one. The truth is we advertise the 
target lead-in price products, knowing that they will often end 
up being the lower sales. 

Sugar grasped the importance of what went on in the shop, yet he 
also believed there was little which a retailer could do to alter the 
law of the market: if Amstrad's products are good, they will sell; if 
not, they won't. 

Amstrad's boss drew the implication that he should never expect 
any favours from a retailer: 

If they've got confidence in a product, they will stack it up, pile 
it high, spend a bit of their own advertising on it. If it starts to 
sell like wildfire, they'll be on the phone telling me what a 
wonderful job they're doing for me and how lucky we are to 
have them as a customer. And if the product drops dead and it 
doesn't sell, they'll tell me that their warehouses are full of 
refrigerators and, sorry, they can't book my van in today. They'll 
tell me not to worry, because they know they gave me an order 
and they will take the stuff eventually. And then they will renege 
on the order. That is the life of a retailer. 



323 



DAVID THOMAS 

Indeed, Sugar formed the view that there was no point in trusting 
a retailer's written order: 

The thing about dealing with a retailer is that their order 
basically is not worth the paper it is written on - unless you 
don't ever want to deal with them again. If you need an ongoing 
relationship with a retailer, there is no way you can start 
banging your hands on the table and insisting on the letter of 
the law. If they don't want to take the stuff, they won't take it. 
The only way you could make them take it is to sue. You'd win 
in court, but you'd never sell them another thing. It's the law of 
survival. 

Sugar carried this bleakly realistic view into his dealings with his 
retail customers. As one large retailer says: 'He had a really 
positive dislike of his customers. He was unsympathetic to the 
retailers' need. The relationship between the buyers and himself 
became very strained. He was aggressive to a quite mind-boggling 
degree. Yet it was all inconsistent with the rather nice guy you 
would speak to ten minutes later, when the venom had been 
poured out and gone.' 

A more common complaint was that Sugar's staff imitated him. 
The disadvantage of the tight Amstrad culture was that it encour- 
aged those caught up in it to copy the ways of their leader. But 
Sugar is inimitable, as one customer explains: They imitated his 
aggressiveness, the worst side of his character, without having the 
intelligence to know that beneath it was a very perceptive, 
humorous, intelligent guy.' 

William Poel, who worked for Amstrad when the PCW8256 was 
taking the market by storm, develops this point: 

Amstrad's sales organization forgot some of the basic tenets of 
salesmanship. Like one day you will find yourself short of 
customers and you will have to revisit those you told to bugger 
off six months ago. They failed to operate their position of 
monopoly in a benign fashion, and were seen as taking a 
malevolent glee in telling people to get stuffed. In the office, 



324 



ALAN sugar: the amstrad story 

there was a kind of school-room camaraderie and people tended 
to get carried away, whereas Mr Sugar knew where to draw the 
line in dealing with a customer. When he told a customer to 
stuff himself, there was a twinkle in his eye and the customer 
knew it was all part of the game. The underlings were unable to 
apply Sugar's perception and wit and kept blundering through, 
saying 'bugger off' without a twinkle in their eye. Customers 
could take any amount of abuse from Mr Sugar, but they 
couldn't take any amount of abuse from the second stringers. 

During the second half of the 1980s, Amstrad's most important 
U.K. customer was Dixons. The market leader in electrical goods, 
Dixons had spearheaded the introduction of many items which 
would not have been found in Britain's high streets at the start of 
the decade - video recorders, telephones, video cameras, Walk- 
mans. The Amstrad-Dixons partnership was crucial in encourag- 
ing the rapid take-up of word processing and personal computing 
by the British public. 

In the aftermath of the PCW8256 and PC 1512 triumphs, 
relations between Amstrad and Dixons became very close. Stanley 
Kalms, Dixons chairman, had demonstrated entrepreneurial flair 
on a par with Sugar's in forging the world's largest electrical goods 
retailers out of the modest passport photograph business he 
inherited from his Polish-born father shortly after the Second 
World War. In his way, he could be almost as abrasive as Sugar. 

Negotiations between the two companies assumed great import- 
ance for both sides. Alan Sugar would invariably lead for Amstrad, 
sitting across the table from Dixons' top executives, Kalms and 
Mark Souhami, the managing director. The flavour of the 
occasions is described by a participant: 

These were not meetings where you would want to talk unless 
you had to because you would get ripped apart - by your own 
side if not by the other. They were fighting over volume and 
percentage points for hours and hours. It was like playing chess, 
only more aggressive. Sugar would say, 'You're going to take 
more than effing 2,000 pieces and you're going to pay more than 



325 



DAVID THOMAS 

this effing price.' Kalms and Souhami would come back at him. 
They're professionals and they can take the verbal. But they 
weren't as rude as him. I always wondered whether that was 
part of his act, because he gained a tremendous advantage 
whenever he broke someone's hard-nosed business exterior. And 
then he would come up with all the reasons and all the angles 
why they should take more of his equipment. It was like role 
playing between Sugar and Kalms. They would slag each other 
off for a couple of hours, then step outside the room and laugh. 

Since Dixons was taking larger volumes than other Amstrad 
customers, it inevitably squeezed a better deal out of the company. 
'Alan had enormous margins. His profitability was quite phenom- 
enal. And if anyone was able to negotiate well, he was. But sure 
we got better margins,' Stanley Kalms says. 

Uniquely among Amstrad customers, Dixons had an arrange- 
ment whereby it would pick up and pay for Amstrad's goods in 
the Far East; it was buying in such large volumes that it could 
justify hiring its own ships. One person privy to Amstrad's deals 
with Dixons estimates that the latter was able to cut 5 per cent off 
Amstrad's prices on the back of this arrangement - more than 
enough to cover the freight and insurance costs which fell on 
Dixons through paying for its own shipping: 'Dixons 5 per cent 
advantage meant a lot. That's a lot. And that was one reason why 
they managed to rip everyone else apart.' 

After the PC 1512 was established in 1987, however, both 
Amstrad and Dixons began to feel uneasy about their mutual 
dependence. Dixons searched for other personal computer manu- 
facturers so that it could diversify its sources of computers. This 
caused growing friction between Amstrad and Dixons, which 
briefly surfaced in March 1988 when the retail chain decided to 
back a new personal computer from Olivetti, the Italian electronics 
company. Sugar publicly attacked Olivetti's computer, the PCI, 
which in the event did not sell well. 

For his part, Sugar realized how important Dixons was to 
Amstrad's success in the U.K. Back in Amstrad headquarters, he 
would jump on staff he heard criticizing the retail group: 'Don't 



326 



ALAN sugar: the amstrad story 

slag off Dixons. They're our bread and butter in the U.K. Don't 
you ever forget it,' he would say. Yet he was also aware of the 
parallel with Amstrad's over-dependence on Comet at the end of 
the 1970s. If Dixons had a bad year, then Amstrad's British sales 
would inevitably suffer. 

Amstrad began a drive to build up sales through independent 
computer dealers. Not only would they reduce Amstrad's depen- 
dence on Dixons, they also offered a better entry into the corporate 
market than Dixons. The wisdom of this strategy became apparent 
in 1989, when Dixons did indeed have a bad year, with profits 
tumbling in the wake of the slowdown in consumer spending. 
Some retail analysts argued that Dixons had been slow to adapt to 
new retailing trends in the second half of the 1980s; the uncertainty 
hanging over the company was compounded at the end of 1989 
when it was the subject of a hostile bid by Kingfisher, the group 
which owns Woolworth and Comet. 

All in all, in 1988 and 1989 Amstrad was faced with the need 
either to establish or to repair its relationships with at least some 
customers in three of its important customer groupings - large 
companies, independent dealers and the high street retailers. This 
was a challenge as critical as more measurable problems such as 
the shortage of memory chips. 

The final blow which fell on Amstrad in 1989 was the growing 
realization that matters were badly awry in some of its new 
overseas subsidiaries. Sugar had been used to running the com- 
pany's overseas operations on a loose rein, a strategy which in the 
mid-1980s paid dividends with its success in France and Spain. 
Occasionally he intervened, as in 1985 when he pressured Marion 
Vannier (head of Amstrad France) to drop the crocodile mascot 
which she had placed at the centre of Amstrad's adverts in France. 
Amstrad's French crocodile was at times a cheeky chappie with a 
big grin, and on occasion a suave gentilhomme in sunglasses. The 
crocodile had helped Amstrad to achieve wide brand recognition 
in France - particularly during the successful launch of the 
CPC464 home computer - much as the petrol company Esso is 
known in Britain through its tiger mascot. But Sugar became 
convinced that crocodiles' bad associations were contributing to 



327 



DAVID THOMAS 

the relative failure of the PCW8256 word processor in France, so 
Vannier reluctantly agreed to drop it and copy the anti-typewriter 
campaign adopted by Amstrad in Britain. French sales of the 
PCW8256 did not improve, however. 

Nevertheless, for the most part Sugar had left Vannier and Jose 
Luis Dominguez (Amstrad's agent in Spain) to their own devices. 
He maintained this policy after 1987 when Amstrad bought 
Indescomp, Dominguez' company, appointed the Spaniard as 
head of Amstrad Spain and established subsidiaries in a range of 
other countries including West Germany and the United States. 
Sugar believed that if he chose entrepreneurial heads for his 
overseas subsidiaries, they should be allowed to run their own 
show. Looking back, he regards this attitude as too lax: 'We 
allowed people too much of a free ship. We were blinded by the 
people who claimed to be entrepreneurs in their markets.' 

Spain and West Germany were the main problem areas. The 
number of Amstrad employees and related costs increased rapidly 
in Spain and, worse still, the Spanish subsidiary failed to sell the 
large amount of stock it ordered. Inventories mounted in West 
Germany too, where the new Amstrad subsidiary misread the 
competitive strengths of Amstrad's computers and pitched its 
products at the top end of the market. 

By the middle of 1989, Sugar realized he would have to step in 
and remove Dominguez. When the Spaniard resigned as executive 
head of Amstrad Spain in October, he sold back to Amstrad 4.5 
million of its shares - more than half the share element of the 
original purchase price of Indescomp - for £1 in total. He also 
paid Amstrad £7 million for Amstrad Spain's headquarters situ- 
ated in an expensive part of Madrid, while Amstrad Spain moved 
into a warehouse complex that was being built for it on the 
outskirts of Madrid. 

Sugar removed the head of the West German subsidiary and 
also of the operation in the United States, where again mounting 
expenses were not justified by the level of sales. The actions he 
took were swift and surgical, but they could not prevent the huge 
increase in stock which emerged as one of the main financial 
pressures on Amstrad in the second half of 1989. 



328 



ALAN sugar: the amstrad story 

The problems facing Amstrad combined to make 1989 the worst 
year in the company's history as a public company. On 20 April, 
Amstrad issued a warning as the London Stock Exchange opened 
that delays in introducing the PC2000 series would reduce profits 
for 1988-89 below market expectations. The market responded by 
wiping more than 15 per cent off Amstrad's value: its shares 
plummeted 21 Vfcp to 1 19 ! /2p. A string of gloomy circulars from City 
analysts over the next few months continued to push the share price 
down and by mid-October - a week before the announcement of the 
1988-89 results- Amstrad's stock had fallen to 57p. 

The extent of Amstrad's difficulties was revealed later in 
October when Sugar announced 1988-89 pre-tax profits down 52 
per cent to £76.6 million, on sales barely changed at £626.3 
million. The accounts highlighted the geographical black spots: in 
Spain sales were down a quarter to £81 million, while in Germany 
they had fallen 4 per cent to £28.7 million. But perhaps two figures 
buried in the small print painted the clearest picture of Amstrad's 
troubles. The weak control over foreign subsidiaries, coupled with 
the PC2000 problems, had generated massive inventories: stocks 
now stood at £325 million, roughly six months' production. This 
stock mountain, together with the costs of buying the stake in 
Micron and establishing Amstrad's overseas subsidiaries, had 
reversed Amstrad's traditionally cash positive position: the com- 
pany now had borrowings of more than £100 million. 

The one aggressive note which Sugar sounded was to warn that 
he might take Amstrad private again if the City did not improve 
the rating of its shares. Sugar announced that at the annual 
general meeting in December Amstrad's board would seek author- 
ity to buy its own shares. He pointed out that since the share price 
valued Amstrad at less than its net asset value, there seemed little 
point in being a public company: 'If the market does not like us as 
a company and reflects that in its share price, and if the share 
price remains low, we will buy it,' he said. 

Sugar also declared in his chairman's statement that his imme- 
diate priority would be to sort out the inventories and the cash 
position: 'The short-term objective of getting the company in good 
shape will focus on securing its existing assets and its good brand 



329* 



DAVID THOMAS 



name position in the market. It may be a bitter pill to swallow for 
some investors, since this policy may be at the expense of chasing 
profits, but [it] will bring back the good liquidity which has been 
a feature of Amstrad's balance sheet for the last three years.' 

This clear warning that higher profits were not on the cards in 
the short term prompted Amstrad's shares to fall another 4p. Its 
shares stood at 47p, less than a quarter of the price they had 
commanded a year previously. 

In December, about 200 small shareholders gathered for the 
A.G.M. at the Tower Hotel down river from the City. The 
adulation which many seemed to feel for Sugar was still apparent, 
but it was overlaid by an air of anxiety about what would happen 
to Amstrad. Sugar was in emollient mood, more or less ruling out 
the idea of taking Amstrad private in the immediate future. When 
asked by a shareholder whether he intended to emulate Richard 
Branson, who had recently re-privatized his company Virgin, he 
replied, 'I'm not going up in a balloon, if that's what you mean.' 

Indeed, Sugar got as close as he had ever done to complimenting 
the City: 'Contrary to what people think, I and my directors can't 
complain about being a public company,' he said. 'Rebuilding our 
image in the stock market means rebuilding our profits. Nothing 
else.' 



330 



• 17- 
Controlling the Tiger 



One common thread ran through almost all the problems which 
beset Amstrad in 1988 and 1989 - the lack of management control. 
In the mid-1980s, as Lord Young put it, Alan Sugar had been 
riding a tiger; the tiger was growth on a scale that few of the 
world's businesses ever succeed in generating. During the first 
seven years of the 1980s, the annual increase in Amstrad's sales 
was never less than 56 per cent and usually substantially more. In 
1986 it was already a large company, yet sales grew by an 
astonishing 123 per cent. Sugar's overwhelming priority had been 
to feed the tiger, creating the conditions which would continue to 
fuel this heady expansion. In Amstrad's case, this meant devising 
innovative new products and marketing them aggressively. 

However, there comes a point in the life of every fast-growing 
business when a change of gear is forced on the founding entre- 
preneur. A crisis - costs running ahead of sales, an over-optimistic 
build-up of stocks, a downturn in the market - brings home to the 
founder what in retrospect appears so obvious: that elementary 
management controls, the good housekeeping rules of a business, 
have been neglected during the years of heroic growth. The 
priority then switches from feeding the tiger to controlling it, and 
the task is urgent: failure could result in the tiger devouring the 
business. 

The moment when it becomes time to control the tiger varies 
from business to business and from entrepreneur to entrepreneur. 



331 



DAVID THOMAS 



Some companies never reach this point, because they do not go 
through a phase of hyper-growth. Other entrepreneurs, by contrast 
fee! themselves losing control almost as soon as their company has 
pulled itself above the massed ranks of small businesses. 

There appears to be some kind of pattern among that tiny 
handful of businesses which become world-class operations from a 
standing start. Often they threaten to run out of control when they 
pass through the $1 billion turnover barrier, and it was at about 
this point that Amstrad stumbled. This was also the stage, for 
example, at which Apple Computer lurched into crisis in 1983-84 
- a crisis which sent shock waves through the entrepreneurial 
spawning ground of Silicon Valley. Apple had to confront huge 
inventories, its first-ever quarterly loss and the unpleasant necess- 
ity of dismissing a fifth of its work-force. 

The entrepreneur who created the company often turns out to 
be the wrong person to cope with the crisis. Feeding and control- 
ling the tiger require different skills. The founding entrepreneur is 
often a charismatic visionary driven by the need to develop one 
or two brilliant product ideas - a character unsuited to the 
tedious business of bringing the company under control. The task 
of dealing with Apple's 1983-84 crisis fell to a method man from 
corporate America, John Sculley, a former president of Pepsi- 
Cola. Sculley could not stabilize Apple until he had forced out 
Apple's co-founder, Steve Jobs, the guru who had fired the 
company's early growth with his vision of a computer on every 
desk. 

Alan Sugar is not an entrepreneur in the Steve Jobs mould; he 
has never been driven by the vision of introducing a particular 
product or set of products. On the contrary, from his first 
encounters with the City in the late 1970s he made plain his 
overwhelming priorities: profit and the bottom line. Launching 
new products is of course essential to Amstrad's ability to turn a 
profit, and Amstrad's most successful product ideas have flowed 
from Sugar's brain. But the Tower System, the word processor 
and the I.B.M.-clone were means to an end - profits. And Sugar 
is equally obsessed with the other aspects of business which 
contribute to that end. Sorting through the rushes of Amstrad's 



332- 



ALAN sugar: the amstrad story 

latest advert, managing Amstrad's currency requirements, screw- 
ing the last few cents out of Japanese sub-contractors, negotiating 
an order for half-a-million satellite dishes with Rupert Murdoch - 
he is as ready to turn his mind to these as to thinking about the 
company's next blockbuster product. 

Unlike many entrepreneurs who are driven by one or two 
product ideas, Amstrad's boss is an unusually complete business 
man. Yes, he thinks about new Amstrad products, but he then 
applies his mind to every step which is needed to bring them to 
market, from component purchasing through to the advertising 
campaign. In addition he has a quality which is essential for 
controlling the tiger - a lack of sentiment. 

When the crisis struck, one of Sugar's first decisions was to pull 
Amstrad out of the audio market. By 1989 sales of audio equip- 
ment were contributing only £21.4 million to Amstrad's turnover, 
less than 4 per cent of the total. The crucial factor in prompting 
his decision was that audio was a low profit margin business 
compared with Amstrad's other lines. No matter that audio was 
the business which sustained Amstrad through its first 15 years: 
abandon it. 'There's no margin left in the product and we're just 
wasting our time chasing it,' Sugar said when announcing the 
decision in October 1989. 

Pulling out of audio meant closing Amstrad's assembly line in 
Shoeburyness, with a loss of about 150 jobs. Sugar also closed 
Amstrad's printer factory in Hong Kong and slashed staff numbers 
there. By the end of 1989 Amstrad employed 71 people in Hong 
Kong, down from a peak of 428 a year earlier. He also cut back 
the numbers of Amstrad employees in other subsidiaries such as 
Spain and the United States. 

Cost-cutting was the way to pare back Amstrad's overheads, 
but Sugar also needed to shift the large stocks which the company 
had built up, Amstrad lopped off between £500 and £700 from the 
price of the PC2286 and PC2386 computers when it announced 
the recall of faulty machines in July 1989, helping to restore the 
price competitiveness which had been lost during the delays in 
introducing the machines. In September, during the run-up to 
Christmas, Amstrad launched a promotion and price-cutting 



333 



DAVID THOMAS 

package which included offering a free portable phone with every 
Amstrad portable computer. The company boasted at the time: 
'You could say this is almost a free portable computer with your 
mobile phone.' 

Amstrad also slashed the price of its camcorder, the video 
camera which it had launched in May 1988. Although the £499 
price tag on this machine was lower than anything else on the 
market, it had never sold in the volumes Amstrad had hoped. 
Some observers said this was because Amstrad's low-cost machine 
did not have enough features and lacked facilities such as zoom; 
others argued that most British consumers seemed uninterested in 
the camcorder concept, unlike their counterparts in the United 
States and Japan. Whatever the reasons, Sugar decided it was 
time to shift Amstrad's large camcorder stocks, even if it meant 
disposing of them at or below cost. Dixons sold Amstrad camcor- 
ders at £299 each during the Christmas season: 'At this price, it's 
selling like hot cakes. We're making a good margin. He's not,' 
stated Stanley Kalms, Dixons' chairman. 

These are examples of the immediate measures which Sugar 
took to pull the company round. Yet securing Amstrad's future 
also meant changing its operating methods, a point he fully 
recognized. At the end of 1989 he said: 

There was no financial policing within the organization. There 
was no proper forecasting of overseas subsidiaries. There was no 
proper inventory control. There was no real head-count plan. It 
was a phenomenon brought about by the fact that we had been 
hot-selling the products. As many as we made, we could sell 
them. When you're on that kind of gi-normous growth trend, 
you tend to forget about the instinctive things you used to worry 
about in a much smaller business - looking after the pennies, 
making sure you're not spending too much, that your expenses 
are not running away, that you're not employing too many. 

But cutting out unwanted functions had to go hand-in-hand with 
expanding facets of Amstrad which were under-resourced. 'We've 



334 



ALAN sugar: the amstrad story 

had extravagance in certain areas and we've been stingy in others,' 
noted Sugar. 

Sugar identified senior management as one area for expansion, 
since the company was too large to be run with its previous ultra- 
lean organization: 'We have had such phenomenal growth over 
the past few years that we are still operating with a management 
team you would expect to be running a company a fraction of our 
size/ he told the Sunday Times in April 1989. 

In October Amstrad appointed two new directors-designate to 
its board: Peter Thorns, who joined the company as group finance 
director from Gillette; and John Benjamin, who joined as group 
manufacturing director from the Mars Corporation. Just as 
important as this influx of resources at top level was the reorgani- 
zation which it facilitated. Amstrad's top management was given 
a group responsibility, while a distinct subsidiary to handle day- 
to-day U.K. operations was created for the first time. Barry Young 
- previously in charge of Amstrad's distribution to independent 
retailers - was made U.K. managing director. This structure was 
belated recognition of the fact that Britain accounted for less than 
50 per cent of Amstrad's sales. It also replicated the organization 
that is standard in many large multinationals: national operating 
subsidiaries reporting to group headquarters. Simultaneously, 
Sugar boosted the management in Amstrad's overseas subsidi- 
aries, ensuring that each had a strong financial and a strong 
marketing manager. 'The flamboyance of the salesman will be 
balanced by the financial half,' he said. 

This structure allowed greater oversight of its overseas subsidi- 
aries by Amstrad's headquarters, which the 1989 crisis had shown 
to be necessary. Amstrad's top management, including Sugar, 
devoted one day each week exclusively to a particular foreign 
subsidiary. They also adopted what they called a 'one month on, 
one month off' routine: one month, Amstrad's top team would 
travel to, say, Paris to discuss progress at its French subsidiary; 
the next month, Amstrad France's senior executives would travel 
to Brentwood to do likewise. The discussions range widely over 
sales, stock levels, expenses, advertising, successes, failures, pro- 
motions and so on. Thus, when a foreign subsidiary says that it 

• 335 • 



DAVID THOMAS 

wants to order so many thousand of the PC2000 series, it now has 
to justify this decision before the group team. Malcolm Miller, 
group marketing director, describes this as the 'third stage' of 
Amstrad's evolution: 

The first stage was to get up and run with the product and sell 
it in great volume. The second stage was becoming more 
professional, with more marketing skills, more people, more 
product development. The third stage - now that Amstrad has 
got so big - is to try to understand our business globally. We 
need to have a broad global overview of all our products and of 
how we manage them. We cannot just pick out a single product, 
run with it and forget everything else any longer. 

Hand-in-hand with this more settled structure went a strengthen- 
ing of financial controls within Amstrad. Ken Ashcroft, who 
became corporate finance director in the new organization, had 
predicted at the time when the overseas subsidiaries were estab- 
lished that they would severely affect the company's cash position. 
This was because Amstrad would not receive payment for its 
products until much later in their journey to market. For example, 
when Amstrad's business in Spain was run through the distributor, 
Indescomp, it was paid for an order as soon as it was placed by 
Indescomp. But once Amstrad Spain was established, Amstrad 
was not paid for sales in Spain until payments were received from 
its customers in Spain, most of whom would expect a period of 
credit. Therefore Amstrad had to finance its stock while it was on 
the high seas, in Spanish warehouses and on retailers' shelves. 

Although Ashcroft had anticipated the impact of this change on 
Amstrad's cash flow, no one had anticipated the implication of lax 
controls over the subsidiaries for its inventory levels. Information 
on inventory levels existed within Amstrad, but it was neither 
integrated into a broad picture of worldwide stock positions nor 
considered an essential tool for top management. When Sugar 
realized in early 1989 what was amiss in the subsidiaries, he put 
much greater emphasis on inventory reporting, as Ashcroft 
describes: 'Alan realized that unless he got stuck in himself and 



336 



ALAN sugar: the amstrad story 

made it clear to everyone in the group that inventory control was 
the name of the game, we were going to lose all our cash. So he 
sent out some massive signals to the managing directors, telling 
them that he was starting to focus on inventory totals. He told 
them, "We must have cash control. Therefore we must have 
inventory reporting."' 

By the middle of 1989, Ashcroft had put an integrated stock 
reporting system on to Amstrad's worldwide internal computer 
network. Management in Hong Kong or Dallas would regularly 
key its stock position into the computer. This information could 
then be read in both unit and value terms by Amstrad head- 
quarters. '1989 was the year in which information in the company 
was really addressed. People have realized that they have to run 
the company by numbers. Alan's not a paper man; he doesn't like 
reports. But he now likes inventory reports and he's learned to use 
them,' says Ashcroft. 

To ensure tight control of stock, Amstrad's chairman appointed 
a group inventory controller, Gordon Marshall. Sugar removed 
the power to order new products from the operating subsidiaries 
and invested it in a troika at Amstrad headquarters: Malcolm 
Miller (marketing), John Benjamin (manufacturing) and Gordon 
Marshall (inventory). The company also centred its Continental 
distribution network on two warehouses in Rotterdam and Ham- 
burg, which allowed it to switch goods between countries more 
flexibly as demand dictated. 'We're centralizing our distribution. 
We're not having inventory sitting in ten damn countries all over 
the world,' Sugar said. 

He increased the company's research and development staff, 
placing a special emphasis on testing and checking facilities in a 
bid to avoid any repeat of the PC2000 problems. 'The difference 
between our procedures now and what they were is the difference 
between light and day. We now have a qualification process in 
place. Every piece of software will be tested thoroughly before it is 
shipped.' 

Sugar stressed that Amstrad was now carrying out 'all the 
mundane and boring tasks that you have to do in a high technology 
business. Anyone in the computer industry reading this will say, 



337 



DAVID THOMAS 

"It's boring. It's what all companies should do." Well, we're doing 
it now.' 

Sugar was importing into Amstrad the 'boring' systems which 
most big multinationals would expect to have as a matter of 
course, yet the crisis also prompted changes in core areas of the 
firm's strategy. 

One of these concerned manufacturing. Sugar had begun to 
move some manufacturing requirements from the Far East to 
Europe when the European Commission tightened the anti-dump- 
ing screw in the second half of the 1980s. The Funai-Amstrad joint 
venture in video recorders at Shoeburyness was a response to this 
pressure. It was not a stunning success and Sugar contemplated 
the prospect of moving even more manufacturing to Europe with 
little enthusiasm. In 1988 he confided in Arnold Weinstock, 
managing director of the General Electric Company, that he could 
see little alternative to reducing his dependence on foreign sources 
of supply. He asked whether G.E.C. would be able to make some 
of Amstrad's new PC2000 series. 

In February 1989 Amstrad announced that up to 20 per cent of 
Amstrad's PC2000 range - about 10,000 a month - would be 
made at a Scottish plant in Kirkcaldy, Fife. The deal saved the 
factory, which employed 700 workers, from closure. At that stage 
the plant was jointly run by G.E.C. and Plessey, Lord Weinstock 
having personally intervened with Plessey to ensure that the 
Amstrad contract went to Kirkcaldy. 

Sugar was impressed by the fact that it would cost Amstrad 
only about £10 more per machine to have his computers made in 
Scotland than in the Far East. Most of the cost of sophisticated 
computers like those in the PC2000 range lies in their components, 
not in the labour used to assemble them, so the traditional cheap 
labour advantages of the Far East were becoming less relevant. 
For example, in February 1989 each PC2386 computer contained 
$720 worth of DRAMs. Moreover, more of the key components for 
a computer were now available nearer to home. By 1989 Amstrad 
could buy DRAMs in Europe, because Texas Instruments and 
Siemens had begun to manufacture them there, which allowed 
Amstrad to avoid the 18 per cent duty imposed at the time by the 



338 



ALAN sugar: the amstrad story 

European Commission on chips imported from the United States 
or the Far East. 

The cost advantage of manufacturing sophisticated computers 
in the Far East was therefore rapidly disappearing, while produc- 
ing in Europe yielded two inestimable benefits. First, it avoided 
the risk of anti-dumping actions. Second, it shortened Amstrad's 
supply lines. Now that Amstrad needed to manufacture in such 
large volumes, it was a major disadvantage to have huge stocks on 
the high seas for a couple of months travelling between the Far 
East and European ports. Such long lines of supply constrained 
Amstrad's ability to respond flexibly to changes in customer 
demand - one of the main contributory factors to the increase in 
stocks in 1989. 'It physically costs us £10 more [to make the 
computers in Scotland], but in practical terms it costs us a lot less 
grief and aggravation,' Sugar said. 

He followed the G.E.C. deal by announcing in November 1989 
that S.T.C., another large British electronics group, would make 
Amstrad's fax machines at its factory in Northern Ireland. In 
February 1990, Sugar said that while 25 per cent of Amstrad's 
goods were now made in the European Community and 75 per 
cent elsewhere, he envisaged moving rapidly to a ratio of 60:40 in 
Europe's favour: 'The Far East has no advantage any more in 
computers, because in the past the most labour-intensive part was 
the building and testing of the main printed circuit board. Now 
with automation and modern surface mount technology, that does 
not require so much labour.' 

But Sugar also indicated that British manufacturers would have 
to be sufficiently flexible to switch their production quickly so as 
to meet changes in customer demand: 

We'll make more and more computers here in the U.K. But I 
will insist that the manufacturers become more flexible so that 
they can make the computers in whatever configuration is 
needed. The configurations are changing so much. One moment 
there's a mad demand for hard drive versions, then for single 
drive versions and then for double drive versions. You can never 
get it right. 



339 



DAVID THOMAS 

Customer relations were the other core aspect of Amstrad's 
operations to receive attention when the problems emerged in 
1988 and 1989. The company's customers had never been left to 
fend totally for themselves if a fault developed in their machines. 
Amstrad had a long-standing relationship with Dictaphone, a 
service organization, by which Dictaphone offered on-site repair 
contracts to any end user who wanted one. Most large corporate 
users of Amstrad computers buy their machines from big computer 
distributors, which typically have their own well-staffed repair and 
training organizations. 

However Jim Rice, the responsible Amstrad director, had been 
keen throughout 1988 to improve Amstrad's service and spare 
parts organization. Painfully aware that the existing set-up in 
Shoeburyness was inadequate, he spent almost a year looking for 
a separate site for these functions. He finally bought a 111,000 
square-ft building in Harlow, Essex. Amstrad occupied the Harlow 
premises in May 1989 and rapidly built up the work-force. There 
are still lots of things to be done, but we've got more people, more 
room, better systems and more spares. We don't have the same 
problems as we did in Shoeburyness, so we're able to give a much 
better service,' Rice says. 

At the end of 1989, Amstrad had about twenty people at Harlow 
dedicated to answering telephone queries from end users. There 
were also separate lines for computer dealers, who are more likely to 
be seeking specific spare parts. Rice acknowledged that the Harlow 
operation was a departure from Amstrad's box-shifting culture: T 
see it as a very important aspect of our future strategy, and it's going 
to be developed much more than it currently is,' he stated. 

Meanwhile, Amstrad had been taking steps to differentiate its 
product line. In May 1988 the company paid Swraj Paul's Caparo 
Industries £3.1 million for the Fidelity name, finally capturing a 
quarry which Sugar had first stalked in 1984. Since Paul had 
recently announced the decision to close Fidelity's West London 
factory, all that was left to acquire was the name, which Amstrad 
wanted in order to project its products under different brands. 
Fidelity (or Amstrad Fidelity, initially) would be reserved for the 



340 



ALAN sugar: the amstrad story 

company's leisure products such as video recorders; Sinclair for 
home computers; and Amstrad for business computers and asso- 
ciated products such as printers. 'It's too much to expect the 
consumer to differentiate such a wide range of goods under one 
brand name,' Malcolm Miller said at the time. 

Miller was working on a strategy to improve the ability of 
Amstrad's dealer network to handle the company's most sophisti- 
cated computers. In September 1988, Amstrad announced plans 
to divide the 3,000 computer dealers handling its products into 
two tiers. The top tier of about 1,000 dealers - to be called 
Advanced Systems Centres - would concentrate on the PC2000 
series; this group would be required to dedicate a section of their 
showroom to the series and would also have to offer end users 
installation, testing, delivery and training services, an on-site 
maintenance contract and a technical hotline. 

The appointment of Barry Young in October 1989 to be the 
managing director of Amstrad's new U.K. subsidiary also signalled 
a wish on the company's part to improve its customer relations. 
Young had previously headed Amstrad Distribution Limited 
(A.D.L.) - an operation designed to service the needs of the 
independent retailers - which is based in Stoke-on-Trent on the site 
of Lyons-Turner Enterprises, the wholesaler bought in 1984 when 
Amstrad made its first acquisition. During the three years he had 
been running it, Young had built up A.D.L. into a sophisticated and 
slick operation which supplied a range of services such as display 
material and training to the independent retailers. He was most 
proud of A.D.L. 's ability to offer a next-day delivery service for any 
order, however small, from a dealer anywhere in the U.K. A.D.L. 
claimed 99.7 per cent reliability for this service, which allowed the 
small retailers to keep their stock to a minimum. It required careful 
links between A.D.L.'s computer-based order-taking system and its 
huge warehouse, which daily despatched a constant stream of lorries 
to distribution depots all over Britain. In his new post as Amstrad's 
U.K. boss, Young said, he intended to inject this kind of customer 
consciousness into all Amstrad's British operations. 

The first signs that Sugar's policies to turn Amstrad round were 
working came in February 1990, when the company announced 



341 



DAVID THOMAS 

profits of £30.1 million for the six months to December 1989. 
These results were way down on the £75.3 million achieved in the 
corresponding period of 1988, but well above what the City had 
been expecting. Sales actually increased 7.2 per cent to £373.7 
million. Amstrad's revenues could have been even higher, since 
many of its computer lines were sold out before Christmas, but 
Sugar had not wanted to tempt fate by stockpiling even more 
inventory: 'It is evident and very encouraging that our brand 
image remains untarnished,' he said. 

Sugar also reported on the introduction of more management 
controls into Amstrad: 'We now have in place good systems, and 
although they still have some way to go before we can claim 
perfection, I am confident that we are now able to "keep the score" 
without allowing too much bureaucracy to slow us down,' he wrote 
in his chairman's statement accompanying the half-year results. 

The underlying fundamentals which most concerned the City 
were also encouraging. Sugar disclosed that borrowings had 
peaked at £1 14 million in September 1989, and would be down to 
zero by the end of February 1990, though this is traditionally a 
time of the year when the cash demands on Amstrad are not great. 
Stocks peaked at £335 million in September, and would be down 
a third by the end of February. The main remaining black spot 
was the very top end of Amstrad's range, the PC2386, where the 
company still had two years' worth of stock. 

The decision to sell audio equipment, video recorders and 
camcorders at or below cost over the Christmas season had 
depressed Amstrad's profit margins. But the business was still 
making profits, which was more than could be said for many other 
companies in the computer industry. 

The year 1989 was one of tumbling profits for computer 
companies throughout the world. Shares in Commodore of the 
U.S. plunged 20 per cent in June when it revealed that it was 
making losses. Wang Laboratories, another American computer 
company, was rescued by the banks in August after it had declared 
fourth-quarter losses of $375 million and stopped payments on 
some of its debts. In August, the founder and president of Norsk 
Data, a once high-flying Norwegian computer company, resigned 



•342 



ALAN sugar: the amstrad story 

after warning of heavy losses. Zenith of the U.S. disclosed that it 
was making heavy losses in October, just a few weeks after selling 
its computer arm to Groupe Bull of France. Also in October, 
Italy's Olivetti launched a rights issue after profits in its first half 
plunged 40 per cent. The chairman of Nixdorf of West Germany 
resigned in November after his company notched up losses of DM 
465 million (£277 million) in the first nine months: shortly 
afterwards, Nixdorf was rescued by Siemens. The same month, 
shares in Compaq fell 16 per cent after the U.S. computer 
company issued a profit warning. December saw I.B.M. shed 
10,000 jobs in the U.S., bringing the number of jobs lost in the 
American computer industry during the year to over 40,000. As 
Christmas approached, Apple Computer warned that its seasonal 
sales would be below expectations - a warning which was followed 
by a renewed bout of management restructuring. 

In 1988-89, Sugar's 'year of disaster', Amstrad's profit margins 
fell to 12.2 per cent from the 25 per cent levels they had touched 
during the preceding years. Nevertheless, 12.2 per cent is a margin 
that many computer companies would be pleased to make in a 
good year. Over at G.E.C.'s headquarters in Stanhope Gate, Lord 
Weinstock had been watching Amstrad with interest: 'What's the 
disaster? He is still making money. Yet he's criticized as a failure. 
That's hardly fair.' 



It is impossible to predict what will happen to Amstrad in the 
1990s. Anyone who in 1980 had tried to forecast the company's 
fate over the following decade, including Alan Sugar himself, 
would have been wildly wrong. 

One possibility is that Amstrad will stabilize at its present size 
and then grow steadily but unspectacularly from that base. It 
would hardly be a disgrace for one individual to have created a 
company with annual sales of over £500 million. Yet under this 
scenario, Amstrad's second decade as a public company would be 
less notable and exciting than its first. 

Whether or not Amstrad resumes more vigorous growth will 
turn on its ability to recapture its innovative flair in developing 



343 



DAVID THOMAS 



new products. Sugar recognized this when presenting his half-year 
results in February 1990: 'No matter how many systems and 
controls the company has, it will live or die on its products and 
the manner in which we sell them.' 

By the beginning of the 1990s it was no longer a question - as 
in the Amstrad of old - of looking simply for the one new 
blockbuster product. Amstrad was too big and entrenched in too 
many world markets to think like that any longer. What was 
required was a steady stream of products which could build on its 
existing position. The reorganization carried out by Sugar in 1989 
was designed partly to channel more resources into product devel- 
opment, and also to allow him to devote more time to thinking about 
product ideas himself. As Amstrad entered the 1990s, it was working 
on new products right across its existing lines. 

Yet having said that, it remained true that Amstrad would need 
another breakthrough like the word processor if it was to make a 
further great leap forward. No one inside or outside the company 
can possibly know whether this will happen; it is like asking 
whether Stephen Spielberg will continue to make good films. 

Another possibility is that Amstrad will develop closer links 
with a large British manufacturing group as it gradually moves its 
manufacturing requirements back to Europe. Sugar is clear that 
he does not want Amstrad to become a manufacturer by itself: 
'We will continue to go down the route of not actually manufactur- 
ing much ourselves. We will continue to dream up products, 
develop them, procure some of the major parts, but then get other 
people to make them for us. We will not actually have the factories 
and production lines.' 

But if Amstrad sub-contracts more of its work to a handful of 
British groups, it is possible to envisage the company being drawn 
into a relationship with one of them - a relationship which could 
take weaker or stronger forms. G.E.C.'s Lord Weinstock, for one, 
is thinking along these lines, and says that G.E.C. and Amstrad 
may deepen their collaboration in future. In particular G.E.C. 
could provide the manufacturing and technical infrastructure 
which Amstrad lacks. This in turn could lead to some kind of 
closer relationship between the two companies: 'You could see a 



•344 



ALAN sugar: the amstrad story 

possibility where we together have a factory. He needs technologi- 
cal and production capability. And it's possible that we would 
provide that for him not as something separate, but as part of his 
business. It's possible that one of these days, something like that 
will happen.' 

Weinstock sees a development like that as essential for Amstrad: 
'If he wants to make the business stable, he will have to acquire 
assets. He will have to acquire infrastructure, which comes from 
having a manufacturing and a development capability. He needs 
more than just good ideas now and again. He knows that. He's 
learned a lot in the last couple of years.' 

But Weinstock believes that any future discussions between 
G.E.C. and Amstrad would probably take a different form from 
the talks between the two companies in the mid-1980s: 'It's not 
impossible that we will come together with him in some way. But 
I think now, we would be more likely to concentrate on the things 
he does, so that we help him build up his business, rather than ask 
him to build up part of ours.' 

An Amstrad-G.E.C. collaboration on these lines would provide 
G.E.C. with a launching pad into personal computers and con- 
sumer electronics, areas where it has been absent or weak. 
Weinstock says: 'He offers a possibility of another go. There are 
some technical developments coming along, like high-definition 
television. You can't tell what will happen. But I feel positive 
about that fellow.' 

An Amstrad-G.E.C. joint venture is, of course, just one among 
many possibilities. Another, in theory, is that Alan Sugar will 
throw it all up, retire to his Florida holiday home and live on his 
ample private wealth. Somehow this seems the least likely out- 
come. Sugar is the type of person who quickly becomes bored stiff 
sitting by a pool. Even his private property dealing keeps only a 
small part of his brain engaged. The British business scene has not 
heard the last of Alan Sugar. 



345 



• 18- 
Electric Warrior 



As Amstrad entered the 1990s, Alan Sugar was able to look back 
on a decade of achievements. First - and dearest to his heart - was 
the extraordinary series of results reported by Amstrad in the 
second half of the 1980s. 

Sugar had forged a company which regularly topped the league 
tables of business performance. In both 1987 and 1988, for 
example, Amstrad beat 250 leading British companies to first 
place in the Sunday Times/P-E Inbucon Business Index to emerge 
as Britain's most efficient company. The index measured com- 
panies against three indicators: profit margins, return on assets 
and value added. Similarly, a study of the five years to 1988 by 
James Morrell Associates, an independent research organization, 
found that Amstrad's productivity increases had outstripped those 
in 100 other top British companies. 

Amstrad also shone when compared with international compet- 
itors. Fortune magazine reported that in 1987 only Cray Research, 
a company devoted to the specialist area of supercomputing, had 
a better return on sales out of the 71 computer and electronics 
companies in the Fortune list of 500 top U.S. corporations. 
Amstrad's profit margins were double those of Compaq, widely 
regarded as the most successful manufacturer of LB. M. -compati- 
ble personal computers in the United States. 

There is another way of assessing Sugar's achievements. Look- 
ing back, it is difficult to remember how few European businesses 



346 



ALAN sugar: the amstrad story 

or households boasted personal computers in 1980; the handful of 
adventurous souls who had brought these new gadgets from the 
United States tended to treat them as amusing curiosities, rather 
than as serious business tools. At the start of the 1990s, by 
contrast, it was rare for businesses of any size in the more 
prosperous parts of Europe to lack a computer or word processor. 

Sugar was at the forefront of that business and social revolution 
in large swathes of Europe, particularly in Britain, Spain and 
France. Amstrad transformed the word processor from a product 
whose price had been counted in thousands of pounds to one 
measured in hundreds. Sugar helped to democratize computing 
power, designing and manufacturing computers which were within 
the financial reach of every small business, voluntary organization 
or middle-class family. As a result, by the end of the 1980s in 
Britain he was on his way to achieving the ambition he had 
expressed twenty years earlier as a young man - to make Amstrad 
as famous a brand name as Hoover. Nearly everyone knew what 
was meant by saying that a document had been written 'on my 
Amstrad'. 

In the process Sugar also made himself immensely rich, but the 
wealth he created did not flow to him alone. Amstrad left a trail of 
prosperity in its wake. The creation of the first mass personal 
computing market stimulated a range of subsidiary activities such 
as retailing, software writing, peripherals manufacturing and 
magazine publishing. Countless small businesses, owned by mini- 
Alan Sugars, prospered on the back of Amstrad, 

The company's success might be dismissed as a phenomenon of 
the frothy economic boom of the years leading up to 1988. With 
the British economy expanding by almost 5 per cent a year in real 
terms, faster than any other industrial economy except Japan, how 
could Amstrad help but succeed? 

Like any company which relies on sales in the high street, 
Amstrad was indeed affected by the economic downturn of the 
closing years of the 1980s. But to depict it as the creature of the 
Thatcher boom is to miss a deeper point. Amstrad had ridden out 
previous periods of economic turbulence, such as the deep reces- 
sion in the early 1980s, almost unscathed. This is no accident; it 



347 



DAVID THOMAS 

reflects one of the defining features of Amstrad's product strategy 

- its drive to give value-for-money. When times are hard, a small 
business will not necessarily abandon plans to buy a computer; 
but it might lower its sights, cutting the amount of money it is 
prepared to spend and placing greater stress on value-for-money. 
That is the space occupied by Amstrad. 

Sugar's wider relationship to the 1980s - the Thatcher decade 

- is also more complex than is at first apparent. In one sense, he 
is the archetypal Thatcherite entrepreneur. A man whose rise 
from nowhere coincided with the early years of Mrs Thatcher's 
administration, he embodied the meritocratic, anti-Establishment 
values which distinguished Mrs Thatcher's Government from its 
Conservative predecessors. His origins in one of the poorest parts 
of London fed his distaste for the traditional bastions of privilege. 
'There's a new breed of person coming up - the likely lad. You 
see it in the City and everywhere. It's no longer Mr Heathcote- 
Smythe's son who's getting the jobs,' he said in 1988 to John 
Lloyd, the commentator who described Amstrad's chairman 
as an archetypal representative of a new ruling group, 'the 
Disestablishment' . 

The 'Disestablishment' is marked by a reverence for the creative 
power of the market, even when its workings threaten traditional 
power centres. By that measure Sugar belongs firmly within its 
circle. His open contempt for the City, his reliance on the Far East 
for his products rather than on British industry, his decision to sell 
business computers in high street stores like Dixons - Amstrad's 
history is a tale of Sugar affronting and disrupting one entrenched 
business interest after another. 

Sugar is quick to praise Mrs Thatcher for helping to shake up 
Britain. Yet he has such a cool view of the changes ushered in by 
the Conservative administration that it is difficult to brand him 
simply as a child of the 1980s. He contends that there is nothing 
to stop an entrepreneur in Britain and never has been. After all, 
he points out, many people (including Sugar himself) made plenty 
of money in the 1970s, supposedly the low point in the post-war 
history of British business: 'There were never any barriers for me. 
No form of bureaucracy ever held me back.' He adds, T never 



348 



ALAN sugar: the amstrad story 

noticed any difference under Labour or Conservative,' thus dis- 
missing one of the standard excuses advanced by British business 
for its poor performance in the 1970s. 

What Sugar asks of government is to be left alone. It was the 
interfering nature of the European Commission's anti-dumping 
duties which affronted him as much as anything. In return, he will 
not bother government: the Funai-Amstrad joint venture in Shoe- 
buryness was one of the very few examples of Japanese inward 
investment into Britain which did not seek government financial 
aid. 

It made a good story for the media to bracket Sugar with a host 
of former Thatcherite business stars who had fallen or stumbled 
by the end of the 1980s: George Davies of Next, Tony Berry of 
Blue Arrow, the Saatchi brothers, Sir Phil Harris of Harris 
Queensway, Sir Terence Conran of Storehouse, the group which 
takes in Habitat . . . and Alan Sugar of Amstrad. Yet there were 
two major differences between him and many of his supposed 
companions in grief. First, he was still in full control of his 
company as the 1990s began. Second, there were signs that the 
immediate crisis facing Amstrad - the high level of stocks - was 
responding to treatment. 

Indeed, Amstrad never followed many of the fashions which 
swept British business in the 1980s. It was a notable absentee from 
the game of expansion by acquisition which had underpinned the 
success of many of the fallen stars. Sugar was sceptical about the 
wisdom of radical diversification. Many outsiders were too blinded 
by the boldness of Amstrad's launches into new product lines to 
miss this cautious side of Sugar's character. But it also structured 
his early forays overseas, as well as his approach to the risky 
enterprise of satellite broadcasting: in both cases, he ensured that 
others bore the risks. 

Sugar is irritated by attempts to define the 1980s as a decade of 
Thatcher-inspired entrepreneurs which might - literally and met- 
aphorically — have come to an end: 'An entrepreneur, if there's 
such a thing, is a born schemer and thinker up of things. You 
can't have a decade of them coming to an end. It's like saying 
there's a decade of singers or artists coming to an end.' 



349 



DAVID THOMAS 

Relaxed about the prospect of a change of political climate in 
the 1990s, Amstrad's boss points to the experience of President 
Mitterrand in France as evidence that a Labour government in 
Britain could not set the country on a radical new direction: 

If you use France as a template, you can see that Mitterrand 
has recognized that the policies of the right have to be 
implemented. You can't suddenly tax the rich, or so-called rich, 
much more heavily, because everybody is basically rich now. 
There's no such thing as the ordinary British worker any more. 
Everybody wants their colour TVs, their microwaves and their 
cars. The Left cannot be as our fathers would remember it in 
the old days. If the Labour Government comes in, their policies 
will have to be virtually the same. 

Sugar's business strategy was not a creation of the 1980s. Every- 
thing distinctive about Amstrad - its flexible approach to manu- 
facturing, its insistence on designing simple products which would 
appeal to a broad sweep of consumers, its creation of a mass 
market through undercutting the competition on price - was 
already present in the Amstrad of the 1970s, the small audio 
company unknown to most people. 

The features which have distinguished Amstrad have remained 
constant; so too have the criticisms of it as a business phenomenon. 
Perhaps the most powerful is the charge that Sugar has done little 
for the British economy because almost all his company's products 
have been manufactured in the Far East. 

An easy response to this would be to point to the copies of the 
£25 million and £48 million cheques made out from Amstrad to 
the Inland Revenue which hang in Sugar's office - a fraction of 
the wealth which has flowed into the Exchequer as a result of his 
activities. Lord Young, the former Trade and Industry Secretary, 
puts Sugar's attitude on manufacturing into context: 'Manufactur- 
ing wealth is not the be-all and end-all of life. It's wealth creation 
that counts. That's sometimes done by manufacturing and some- 
times done by service sector and other activities. Alan Sugar has 
enabled the British people to buy goods at better value than they 



350 



ALAN sugar: the amstrad story 

would have otherwise. The more his computers are spread widely, 
the more they create wealth.' 

Critics of Amstrad's manufacturing policy have failed to under- 
stand the modern structure of British, and indeed European, 
industry. The company could not have survived, let alone flour- 
ished, if Sugar had tried to meet all its manufacturing needs in 
Britain. His early experience with the box of components badged 
by British electronics firms, but made in Japan, brought home to 
him that many electronic parts were not available at reasonable 
prices outside the Far East. Although the efficiency of the British 
components industry improved dramatically in the 1980s, he still 
had to look elsewhere for key components for his computers. 
DRAM memory chips, for example, were not made in any 
quantities in Europe for most of the decade. 

By the end of the 1980s, many of the long-established names on 
the British electronics scene had scaled back their manufacturing 
in Britain, other than in protected high cost markets such as 
defence. Thorn E.M.I, had all but abandoned electronics manu- 
facturing. Racal was concentrating on service activities such as 
communications. ICL relied on Fujitsu of Japan to supply the 
integrated circuits which lay at the heart of its 'British' computers. 
Ferranti closed its main civil computer manufacturing plant near 
Manchester, accelerating the flight of the British computer indus- 
try from hardware to software. Apricot sold its computer manufac- 
turing division to Japan's Mitsubishi, leaving it to concentrate on 
software and services. 

Sugar was distinctive in his early recognition of the trend 
towards global manufacturing. One of the first entrepreneurs in 
Europe to understand the implications of the manufacturing 
revolution in the 1960s and 1970s in countries like Japan, he 
realized that it would be impossible to compete in the low end of 
markets such as audio unless he manufactured in the Far East. 
This conviction was reinforced by his admiration for Japanese 
attention to detail, a quality he prizes highly. 'When the Japanese 
don't like the colour of something,' he is fond of saying, 'they don't 
say they want it a little more whitish, as a British company might. 



351 



DAVID THOMAS 

They send you a colour card with 15 shades of white and specify 
exactly which they want. Then they make sure they get it.' 

But, unlike many European executives, he was not blinded by 
Japanese manufacturing prowess. He detected weaknesses in the 
Japanese, as he made clear on two separate occasions in 1988: 
'The Japanese are a programmed people. If the Japanese ever 
become flexible, we'd all be dead.' And again: 'The Japanese will 
not dominate computers the way they dominate hi-fi. . . . They 
have no flair. All of the originality comes from Britain or America.' 

Flexibility and flair: these were the two qualities which helped 
to give Sugar and Amstrad a competitive edge - even over the 
Japanese. They were evident in Sugar's innovative approach to 
the design of Amstrad's products, and to the technology embodied 
in them. 

Many industry experts failed to understand the nature of 
Amstrad's innovations. They were fixated on the notion that 
innovation flowed exclusively from fundamental developments in 
basic components such as microprocessors, rather than from 
thinking through novel product designs. Underlying this attitude 
is a prejudice which is contradicted by virtually the entire history 
of post-war British industry: the prejudice that it is easier to design 
products which will sell in large volumes than to make fundamen- 
tally new scientific or technological breakthroughs. 

Amstrad's boss earned the derision of computer buffs because 
he designed his early computers around a ten-year-old micropro- 
cessor. In turn, he poured scorn on the industry's techno-babble 
by declaring his willingness to design his computers around an 
elastic band as long as it did the job. And it was the job which was 
so innovative: Sugar made products cheaply enough to create the 
first European mass market in business computers. Radical inno- 
vations in both product design and technology were necessary to 
deliver these low costs. Sugar's hallmark was cutting out unnecess- 
ary functions and using a common set of components for those 
that remained. An identical design philosophy can be detected in 
the decisions to run all the elements of the Tower Unit from a 
common power supply, and to run the word processor from a 
customized gate array chip. 



352 



ALAN sugar: the amstrad story 

Also innovative when it came to marketing, Sugar presented his 
goods as capable of doing the straightforward things which people 
like himself would want - because that was what they had been 
designed to do. If that meant selling business computers in the 
high street, for example, then so be it. In an age of specialization, 
he had the gift of being able to think simply. In an era of increasing 
product sophistication, he demystified the markets he entered. In 
the midst of an upsurge in niche retailing, Sugar catered for the 
mass market. 

In its practice Amstrad anticipated by over a decade some of 
the theories which kept management gurus and industry pundits 
in well-paid employment in the late 1980s. They noticed that some 
of the most efficient companies were sub-contracting many of their 
activities, leaving only the central management and control func- 
tions in-house. A whole theory of 'core' and 'peripheral' business 
activities was built around this observation, but it was a policy 
which Amstrad had adopted since the mid-1970s. Sugar saw it as 
the best means to maintain Amstrad's flexibility - and flexibility 
is in turn fundamental to the overriding goal of generating profits, 
as he describes: 

We're a very profit-oriented company and the best way to make 
profits is to be active in an area where no one else is. As soon as 
you show the way, you create competition and competition 
means margin erosion. And as soon as margins are eroded, it's 
time to say 'Sayonara'. We moved out of 14" colour televisions, 
car stereos and CB radios years ago. We moved out of audio 
even though it served us well, because it had come to the point 
where we could not make money out of the damn thing any 
more. So it was just a waste of time. You've got to recognize 
when you're no longer the king of the castle in a certain product 
area because everyone and his brother has jumped on the band- 
waggon. When you wake up on a Monday morning, nobody 
anywhere in the world owes you a living. The minute someone 
offers a lower price or a better service, you will be dropped like 
a hot brick. As long as you recognize that and don't expect any 
kind of loyalty or relationship, you can't go too far wrong. 



353 



DAVID THOMAS 



Flexibility is ingrained in Sugar because from the age of 21 he has 
been engaged in consumer electronics, one of the most competitive 
businesses in the world. Its defining features - high volumes and 
tight margins - force companies into a never-ending race to create 
new products. A hot seller one year is old news the next. In the 
computer industry in the 1990s, this competitive dance will be 
driven by the incredible speed of development of the underlying 
semiconductor technology. Chip product cycles are down to three 
years, with each new generation of chips typically an order more 
powerful than its predecessor. In 1980, for example, the standard 
memory chip was not capable of storing the information contained 
on the page of a book. By 1990, the industry was introducing a 
memory chip capable of storing the whole book and was already 
laying plans for a new chip 16 times more powerful. 

These advances will generate radically improved products in 
personal computing, telecommunications, fax machines, broad- 
casting, cameras, video recorders - and no doubt in areas as yet 
unimagined. But entrepreneurs will have to be flexible and fast 
moving to take advantage of such dazzling developments. 

Alan Sugar has that flexibility and speed of reaction. He learnt 
it not in a business school or out of a textbook, but in the back 
streets of London trading a motley assortment of electronic goods 
as a young man. That experience honed the fighting qualities 
acquired during his upbringing in the East End. In creating 
Amstrad, he had to battle against the keepers of conventional 
wisdom in the City, the media, the European Commission, the 
established electronics giants and almost any other group which 
crossed his path. Sugar did not suffer fools at all. When he thought 
people were wrong he let them know, irrespective of their standing. 
Many returned the compliment by predicting that he would fall 
flat on his face. Proving them wrong is one of the main forces 
driving him forward. 



•354 



ALAN sugar: the amstrad story 
AMSTRAD'S RECORD BEFORE FLOTATION 

1970 1971 1972 1973 1974 1975-76* 1977 1978 1979 

Turnover (£000) 99.6 207.5 726 1327 1689 3594 4033 4469 5978 

Exports (£000) — — — 142 492 656 381 960 1461 

Pre-tax profits 2 24 99 194 176 354 548 650 908 

(£000) 

Key Ratios 

Turnover growth — 108% 250% 83% 27% N/A N/A 11% 34% 

Pre-tax profit — 1100% 313% 96% (9%) N/A N/A 19% 52% 

growth 

Pre-tax profit as a 2% 11.6% 13.6% 14.6% 10.4% 9.8% 13.6% 14.5% 15.2% 

percentage of sales 

* 18 months 



355 



DAVID THOMAS 

AMSTRAD'S RECORD AFTER FLOTATION 

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 

Turnover (£m) 8.8 14.1 28.1 51.8 84.9 136.1 304.1 511.8 625.4 626.3 

Exports (£m) 1.3 0.2 0.1 5.3 10.8 72.8 174.9 289.3 359.7 372.9 

Pre-tax profit 1.4 2.4 4.8 8.0 9.1 20.2 75.3 135.7 160.4 76.6 

(£m) 

Post-tax profit 1.1 1.2 2.6 5.3 5.7 14.0 52.0 93.4 105.1 51.1 

(£m) 

Dividends 0.06 0.11 0.14 0.16 0.19 0.27 0.49 0.97 1.87 1.87 

pence/share 

(incl. tax credit) 

Dividend cover 5 3 6 10 8 14 27 24 14 6 

(times) 

Profit retained 0.8 1.1 2.4 5.1 5.0 12.9 47.2 89.6 93.0 43.1 

for the year 

(£m) 

Earnings pence/ 0.22 0.26 0.54 1.12 1.17 2.57 9.54 17.13 18.99 9.01 
share after tax 

Share capital 3.7 4.8 7.2 12.3 29.2 42.2 88.1 179.5 256.2 310.8 

and reserves 

(£m) 

Key Ratios 

Turnover 56.5% 61.2% 98.6% 84.6% 64.0% 60.2% 123.5% 68.3% 22.2% 0.1% 

growth 

Pre-tax profit 49.9% 74.6% 100.8% 68.6% 13.3% 121.1% 273.5% 80.3% 18.2% (52.2%) 
growth 

Pre-tax profit as 15.5% 16.8% 17.0% 15.5% 10.7% 14.8% 24.8% 26.5% 25.6% 12.2% 

a percentage of 

sales 



356 



Index 



A B Electronic 189, 195 

Acorn 116, 118, 135, 137, 138, 
139, 141, 147, 164; foreign 
market 148, 261; products 119, 
121, 125, 135, 148 

A.E.I. 90 

Aiwa 102 

Alan Sugar Foundation 279 

Alexander, Robert, Lord 106 

Altwasser, Richard 199-203, 254. 
318 

Ambit International 115, 117, 122, 
128, 130, 164-5 

Amritraj, Vijay 278, 279 

Amsoft 164-7 

Amstrad: see also Sugar, Alan 
structure: foundation 40-3; 
expansion 57-8, 79-83, 96-7, 
113, 199,218-19,251,253-6, 
331-2; management 8, 65, 
83-92, 113,238-41,252,325, 
335-7, 342; advertising and 
marketing 4, 89-90, 96, 103, 
177-80, 227-8, 247-8, 319-20; 
workforce 63-4, 85, 1 13, 186, 
238; r and d 199-201,318,337; 



Amsoft, Users' Club and 
servicing 164-5, 232, 316, 
339-40; Amstrad Distribution 
Ltd 341 

business style: value for money 
3-4,69, 124, 127, 159, 165, 
172-3,245-7,320-1,347-8; 
flexibility 58, 101, 111, 187,339, 
350, 353; working methods 128, 
238-41, 244-9, 254-6, 258-9; 
'Amstrad effect' 4, 81, 175, 184, 
185,233, 316; reputed 
unreliability 7, 70-2, 228-31; 
recession periods 93, 96, 137-40 
finance: structure 9, 10, 83, 90, 
107-8, 137, 186,252-3,336-7, 
342; flotation 2, 8-15, 83, 90, 
108; sales and profits (1970s) 
10, 65, 79, 82, 93, 356; profits 
and share prices (1980-5) 93, 
107-8, 109-10, 112, 137, 
139-40,212-13,355; (1985-8) 
182-4, 186, 218, 232-3, 235-6, 
274, 355; (1988-9) 306-7,314, 
329-30, 342, 355; share options 
6, 256-8 



357 



DAVID THOMAS 



subsidiaries: France 65, 97, 99, 
140-6,151-4, 156-9,261,273, 
327-8, 335-6; Spain 140, 
149-51, 154-9,202,261,264, 
273, 328, 335-6; USA 168, 222, 
261-4, 273, 328, 335; West 
Germany 140, 222, 264-7, 273, 
301-2, 313, 328; others 267, 273, 
302 

supply and manufacture: 
concepts 4, 7, 52, 113-14, 
172-3,224,261,312,350-1, 
UK operations 62-4, 82, 293-4, 
312,333,338-41,98-100, 
271-2, see also Shoeburyness 
plant; Japan 58-9, 61,82, 
131-2, 269-70, see also Funai; 
Hong Kong 82, 85, 98-9, 261, 
267-70, 333; other Far East 
countries 61, 82, 100, 131, 177, 
191,268-70,272,311-12, 
338-9; EC constraints 271-4, 
294, 338 

retailing: policy 97, 141, 
322-4, 327; Comet link 50-3, 
65-8, 71, 77-8,83,97,216,327; 
Rumbelows 71, 78, 95, 97, 137; 
Dixons 141, 180-2, 185,231, 
293, 298, 325-7, 334; Currys 33, 
78, 95, 97, 216, 298; Wildings 
185-6, 231; Woolworths 77, 78, 
97; other outlets 97, 167, 231-2, 
284-5, 327; Nigerian deal 
59-61; BPI case 102-7 
joint ventures: Sinclair deal 
189-99; G.E.C. 208-13, 344-5; 
Funai 271, 272-3, 312-13, 338, 
349; Micron 310-1 1; Thorn 
EMI 217-18; other acquisitions 



213-16, 341; diversification 
216-19 

products: cigarette lighters and 
intercoms 40-1; audio 
equipment 44-7, 53-6, 61-2, 
64, 69-70, 81-2, 102, 184, 333; 
Tower System 76-8, 82, 89-90, 
96, 100, 124, 332; in-car radio 
and CB 81-2, 94-6, 100, 115; 
TV and video 100-1, 127, 184; 
CPC464 115-34, 140, 150-6, 
168, 169, 183,228,317,327; 
CPC664 168;CPC6128 156, 
168-9, 178; PCW8256 (Joyce) 
and Ant 156, 161-4, 169-78, 
183, 185-6, 222, 223, 265, 
327-8; Sinclair range 191, 
200-2, 203, 340-1; PC1512 159, 
223-33, 309, 315; PC1640 234, 
236, 242, 274, 309, 319; 
PCW9512 236, 274; PC200 318; 
PC2000 range 313-20, 329, 333, 
337, 338, 342; software 1 18-20, 
130-1, 133-4, 138, 150-1, 175, 
184-5, 224-6; satellite TV 
285-7, 290-305; camcorder 334, 
342; facsimile machine 318, 339; 
portable computer 318, 334 

Anglia Television 285 

Apple, Aaron 21, 29 

Apple, John 26-7 

Apple Computer 135-6, 147, 220, 
221,225,249-50,310,332,343 

Apricot 141,261,351 

Argos 95 

Ashcroft, Ken 195-6, 251-4, 257, 
336-7 

Astra 288, 301-2 

Atari 134, 136, 147 



358 



ALAN sugar: the amstrad story 



Audio Supplies 44 
Audiotronic Holdings 78-80 
Australia: Amstrad subsidiary 267 
Avnet International 311 

Bailey, Paul 133-4,225,258-9 

Bang and Olufsen 247 

Barclays Bank 116-17, 189, 194 

B.B.C. 230 

Belgium: Amstrad subsidiary 267 

Benjamin, John 335, 337 

Berry, Tony 349 

Binatone 34-5, 42, 43, 61,81, 

95-6 
Birds Eye 88, 255 
Blue Arrow 349 
Bond, Alan 287 
Boots 141 

Boston Consulting Group 219, 260 
Branson, Richard 279, 280, 285, 

287, 330 
Brentwood HQ 165-6, 238-9 
Brentwood School 118, 119 
British Phonographic Industry 

(B.P.I.) 102-7 
British Satellite Broadcasting 

(B.S.B.) 285-8,290,291,300, 

304-5 
British Telecom 5, 6 
Brookhouse school 23-6, 27 
BSR47 

Buckley, Sir Denys 105 
Business in the Community 279 

Callen So 99, 269, 270 
Cambridge University 118-19 
Campaign 180 

Caparo Industries 213-14, 340 
CBS 105-6 



Chase Manhattan Securities 

233-5 
China: Amstrad dealings 268, 312 
Chinchen, George 53, 54-5, 85 
Ciba-Geigy 242-3 
Citibank 189 
Citizen's Band Radio (CB) 94-6, 

115, 140,213 
City, the: and Amstrad 137-8, 

139-40, 182-3,212,232-5, 

306-7, 314, 329, 342; Amstrad 

flotation 2, 10-15, 109-12; 

Sugar's relationships 7, 10-13, 

1 1 1-13, 234-5, 244, 304-5, 330, 

348 
City University Business School 

244-9,251,262 
Clayton, Richard 119-21, 132, 

170, 171-2, 173 
Clerical Medical and General Life 

Assurance 11, 14, 110 
Cogel 142-3, 145 
Comet 12, 50-1, 216, 251-2, 327; 

Amstrad link-up 52-3, 65-8, 71, 

77-8,80,83, 137,327 
Commodore 116, 125, 134, 136, 

148, 199, 263, 342 
Compaq 227, 343, 346 
Concentric 293 
Confederation of British Industry 

(CBI) 93, 94 
Conran, Sir Terence 349 
Corte Ingles 158 
Cray Research 346 
Cross, Malcolm and Maureen 

31-2,37,44,55-6,74 
Curry, Chris 138, 139 
Currys33, 95, 97, 216, 298 

Daily Mail 138 

Daily Telegraph 10, 96, 199 



359 



DAVID THOMAS 



Dallas Electrical 44 

Darty 158 

Data Recall 120, 121, 164, 171 

Dataquest221,310 

Davies, George 349 

Delaney, Greg 178-9, 180, 320 

Delaney Fletcher Delaney 1 78-9, 

227, 319 
Dictaphone 231, 316, 340 
Digital Research 133-4, 160, 

225-6, 263 
Dixons 231, 277, 325, 327; 

Amstrad deals 141, 180-2, 185, 

188-90, 325-7, 334; Sky TV 

project 293, 298, 300 
Dominguez, Jose Luis 146-51, 

154-8, 264, 328 
Dragon 138 

Egan, Sir John 280-1 

Electrical Retailer and Trader 1 76-7 

Epson 181 

Eriera, Gerry and Norma 74^5, 

98, 276 
Euromarche 158 
European Commission 7, 101, 

215, 270-3, 294, 338, 349 
Evening Standard 92, 108 

Ferguson 214-16, 247 

Ferranti 126, 130-1, 183, 196, 351 

Fidelity (Radio) 62-4, 71,213, 

214,215,340 
Financial Times 139, 140, 215, 229, 

285, 300; on Amstrad 9-10, 

182-3, 199, 232, 234 
Financial Weekly 10 
Fisher, Howard 119 
Fortune 280, 346 



France: subsidiary set up 97, 99, 
142-6, 156-9, 335-6; sales 65, 
140, 151-4,261,273,327-8; 
politics 350 

Frances (Alan's secretary) 283-4 

Fraser, Alan 242 

Fujitsu 351 

Funai 270, 271, 272-3, 312-13, 
338, 349 

Garrard Ltd 52 

Gates, Bill 221 

General Electrical Company 

(G.E.C.)4, 121,207-12,217, 

338, 344-5; Sky TV project 

293-4, 296, 299 
Gilbert Eliott 110 
Girozentrale 1 1 
Global Audio 44-5, 47, 56 
Goldman, David 184-5 
Gordon, Guy 15, 40 
Granada 284-5 
Grant, John 196-7 
Grattan 97 
Greenwell, Pip 1-2 
Greenwell, W. 1, 5, 10, 12, 14, 

109-10, 111-12,209,212-13 
Groupe Bull 343 
Guardian 107, 134, 167, 227 
G. W. Smith 79 

Habitat 50, 349 

Hackney Gazette 31 

Hall, Chris 1 19-29 passim, 132, 

169-71 
Harlow site 340 
Harris, Jim 22-3, 24, 26 
Harris, Sir Phil 349 
Harvey-Jones, Sir John 280 



360- 



ALAN sugar: the amstrad story 



Henson, John 35 

Henson, Peter 34 

Herbert Smith 12 

Hewer, Nick 134, 198,280 

Hi-Fi News 4\, 55, 69, 70 

Hitachi 162,210 

Holland-Bos worth, Tim 9, 10, 1 1, 

13, 14,80 
Hollingbery, Michael 12, 50-1, 

79_80, 110-11 
Hong Kong: Amstrad subsidiary 

82, 85, 98-9, 257, 261, 267-70, 

312,333 
Horrell, Peter 227, 319 
Hotpoint209, 211-12 
House of Lords 106-7 
Howe, Sir Geoffrey 93 

I.B.A. see Independent 

Broadcasting Authority 
I.B.M. 125, 182,220-4,228-9, 

232-3,249,319,320,321; 

products 126, 133, 226-7, 242, 

320 
I.C.I. 227-9 
ICL 141,351 
Independent Broadcasting 

Authority (I.B.A.) 285-6, 287 
Independent Electrical Retailer 69 
Independent Television 

Companies Association 

(I.T.C.A.) 179,227 
Indescomp (Amstrad Espafia) 

147-8, 149-50, 158,264,328, 

336 
Intel 169,317 

International Herald Tribune 199 
International Management 187 
Ip, Isaac 269 



Isherwood, Patrick 103, 104, 105 
Italy: Amstrad subsidiary 267, 273 
I.T.C.A. see Independent 

Television Companies 

Association 
I.T.T. 58, 300 

Jackson, Peter 137 

Jaguar 280-1 

James Morrell Associates 346 

Japan: electronics industry 5, 9, 

61.81, 102, 181,289,308-9, 
349, 351-2; Amstrad's interests 
58-9,60,61,82, 131-2,269-70; 
and EC 101,270,272-3 

Jobs, Steve 135, 250, 332 
Johnson, David 71, 78, 81, 91 
Jones, Mark-Eric (Mej) 120-32 

passim, 164, 169-70, 172, 174, 

223,224,317-18 
Joseph Priestley school 22-3 
Jost, Helmut 267 
JVC 69, 70, 214, 247 

Kalms, Stanley 180-2, 188, 277, 

325-6, 334 
Kentridge, Sydney 106 
Kewney, Guy 129, 134-5, 176, 

227, 228, 318 
Kieve, Diana and Jeffrey 18, 19, 

20,21,23 
Kingfisher 327 
Kleinwort Benson 9, 10, 12, 15, 

80, 1 1 1 
Korea, South: Amstrad's activities 

61.82, 131, 174, 177,201,268, 
270; and EC 272-3 

L.&N. 62-4, 70, 71,82, 92, 97, 
213 



361 



DAVID THOMAS 



Ladbroke 80 

Laister, Peter 217-18 

Lalvani, Gulu 34, 35, 38, 42, 60-1, 

79, 95-6 
Lasky's 47, 78-9, 80 
Lawrence, Len 62 
Lawrence, Robert 70-1 
Lawton, Lord Justice 104-5 
Lazarus, Barney and Myrna 276 
Le Monde 1 58 
Lettice,John227,321 
Lewin, Colin 41-2, 45, 47 
Lightowler, Nick 61-2, 66, 67-8, 

71,79, 137 
Lloyd, John 348 
Lloyds Bank 83 
Locomotive Software 119, 122, 

\25-9 passim, 132, 164, 169-71, 

173-4,223 
Lotus 133 
Lyons-Turner Enterprises 213, 

215,341 

Mackay, Andrew 198 

Mak, Simon 98-9 

Malaysia: Amstrad manufacture 
312 

Marks, Ronnie 42-3, 58, 72 

Marshall, Gordon 337 

Mason, Gerry 51-3, 66-7, 68, 71, 
77-8 

Matsushita 162 

media: trade press 69, 105; trade 
press and Amstrad 9, 171, 176, 
227, 228, 315; Sugar's 
relationships 9, 69, 72, 89-90, 
235-6, 244, 280, 349; Sinclair 
deal 198-9; Sky TV 298-300 

MEJ Electronics 120, 122, 125-9 



passim, 162, 164,170-2,200, 

223, 224, 316-18 
Merrette, Ted 195 
Michael Joyce Consultants 134 
Micron Technology 309, 310-1 1, 

329 
Microscope magazine 183, 228, 229, 

321 
Microsoft 118, 133, 221, 225, 226 
Miller, Malcolm joins Amstrad 

87-9, 91-2; marketing 

management 89-90, 177-8, 

239-40, 243, 253, 254-6, 257; 

Sinclair deal 195-6; share 

option 257; company reform 

307,313,319,336,337,341 
Ministry of Defence 5, 85-6 
Ministry of Education and Science 

29-31 
Mitterrand, President Francois 

350 
Mitsubishi 351 
Moore, Bobby 278 
Moore, Vernon 263-4 
Mordant Latham 90 
Morgan, Greg 110-11, 112-13 
Morita, Akio 288-9 
Morris, Ashley 43-4, 47-8, 56 
Mullard 58 
Murdoch, Rupert 283-4, 287-93, 

294-5, 296-301, 303-4 
Myles, Howard 10,209,213 

Nationale-Nederlanden 147 
Netherlands: Amstrad subsidiary 

267, 302 
News Corporation 304 
News International 287-8, 289, 

295 



362 



ALAN sugar: the amstrad story 



Next 50, 349 

Nicholls, Lord Justice 105 

Nigeria: Amstrad deal 59-61 

Nine Tiles Information 197 

Nixdorf310, 343 

Nixon, Ron 62-4, 71,92, 97 

Norsk Data 342 

Observer 227 

Olivetti 139,227,229,326,343 

Ord-Hume, Arthur 69, 70, 106 

Oricll6, 139, 151-2 

Orion 61-2, 127, 131-2, 162,268 

P. &P. 316 

P.A. Technology 200 

Paul, Swraj 213,214, 340 

PC Dealer 228, 229, 318 

PC User 223 y 242, 315 

Pearson group 285 

PepsiCo 250 

Perry, Roland: Ambit project 

115-22, 127, 128, 129, 130, 131, 

132; Amsoft 164-5; project 

manager 167-8, 174, 225, 265, 

318 
Personal Computer World 134-5, 

169, 176,227,315 
Philips (of Holland) 68, 90, 184, 

215,247,271, 172 
Pioneer 81 

Plessey58, 59, 217, 338 
Poel, William 115, 117, 122, 130, 

164-5, 166-7, 169,324-5 
Poland: Amstrad market 267 
Pollard, Vic 24, 26 
Popular Computing 1 76 
Power, Thomas 239-43, 277-8 
Premier Radio 42-3, 47, 58 



press see media 
Prince of Wales 279 
Prism 139 

Racal351 

Randall, Stan 85, 98-9, 115, 

267-9,312 
R. Henson Ltd 33-4, 35 
Rice, Jim 13, 90-1, 195-6, 253, 

257,286,312,321,340 
Richard Thomas & Baldwin 31, 

32 
Robinson, Leigh 184 
Robuk Electrical 33 
Romtec 232-3 
Rose, Geoffrey 79, 80 
Rowland, John 139 
Rumbelows 71, 78, 81, 91, 95, 96, 

97, 137 
Rupert Lovell Curtis 178 

Saatchi brothers 349 

Sage Group (Sagesoft) 184-5 

Samsung 310, 311 

Sanyo 181 

Scandinavia: Amstrad hopes 302 

Schneider, Bernhard 264, 266 

Schneider Rundfunkwerke 222, 

264-6,313 
Schofieldjack 134, 167 
Sculleyjohn 249-50, 332 
Sears W'orld Trade (Sears, 

Roebuck) 262-3 
SGS 131 
Sharp 81 

Shearman, Nevill 90 
Shingler, Mike 227 
Shoeburyness plant 98, 99-101, 

127, 137, 166,271-2,321,333, 

338, 349 



363 



DAVID THOMAS 



Shomei Trading 59, 60 

Siemens 31 1,338, 343 

Simons, Ann see Sugar, Ann 

Simons, Johnny and family 37, 38, 
53 

Sinclair, SirClive41, 138, 191-2, 
201, 203, 204-6; home 
computer revolution 14, 136, 
141, 204; Amstrad deal 139, 
190, 191-5, 198-9, 202 

Sinclair Research 116, 118, 135, 
136-7, 139, 147, 164,203,261; 
Amstrad's deal 188-99; 
products 126, 136, 148, 189, 
190-1,200-3,295,340-1 

Sky Television 288-304 

Smith, Peter 288-9, 296-7, 302 

Sony 288-9 

Souhami, Mark 188, 325-6 

South America: Amstrad market 
273 

Southend plant 82, 100 

Spain: Indescomp link 140, 141, 
143, 148-51, 154-8,264,336; 
Amstrad Spain 157, 202, 261, 
273, 328, 333, 335-6 

Stanford University 1 18 

S.T.C. 217, 339 

Stock Market see City, the 

Storehouse 349 

Sugar, Alan: see also Amstrad 
personal life: background and 
relations with parents 2, 5, 
15-27, 37, 39, 84, 278, 348; 
appearance 2, 24, 69; shyness 
and aggression 13, 17-18, 24, 
48-9, 63, 72, 97, 280-2, no 
respecter of persons 7, 68-9, 
104, 131,211,250-1,348,354, 



240-2, see also City; media; 
marriage and family 28, 36-8, 
43, 276-8; lifestyle 6-7, 72-5, 
109, 194, 258-9, 275-7; flying 
and tennis 72-3, 75, 277, 278-9; 
BiC, charities and Gov. 
activities 279-82; property 
interests 274-5, 345; personal 
equity and wealth 3, 8, 108-9, 
137, 183, 213, 274-5, 314, 347 
business methods: profit 
motive 10-1 1,65-6, 192,241, 
245-9, 332; selling and 
marketing flair 3, 25-6, 29, 
31-2, 56, 96, 205-6, 331, 352-3; 
'truck driver' axiom 51, 57, 65, 
78, 165, 315; simplicity and low 
price 47, 77, 89-90, 124, 164, 
182, 204, 243, 350, 352; 
flexibility 3-5, 11,80-2,86-7, 
94, 101, 107, 187,224,350-3; 
retailing logic 322-4; 
negotiating techniques 43, 66-8, 
81, 97, 154-5, 170, 258-9; 'non- 
technical way' 64-5, 70-1, 
122-7, 129-30, 162-3, 203-6; 
managers and management 65, 
83-4, 178-9, 186, 204, 238-43, 
252-3, 324-5, 332-3, 348-9; 
chairman's statements 101, 
113-14, 187,235,236-7,260-1, 
329-30; and workforce 6, 84, 
113, 178, 238-46, acquisitions 
and diversification 214-18, 
344-5; see also Sinclair deal 
below; measures in recession 
93-4, 139-40; 'year of disaster' 
and restructure 8, 307, 309-14, 
329-30, 333-9, 342, 344; 



364 



ALAN sugar: the amstrad story 



analysis and prospects 4-5, 8, 
345-54 

career: early enterprises and 
jobs 2, 24-7, 29-35.40-1; 
enters audio market 41-8. 53-6. 
61-2, 64; Tower System success 
76-8, 96. 100. 107; Comet link- 
up 50-3. 65-8. 77-8. 80, 137; 
Dixons partnership 180-2, 
188-9, 325-7; other retailers 33, 
91, 141, 185-6, 284; Far East 
connection 58-9, 61-2, 98, 100, 
269-70,311-12,333,338, 
351-2; Nigerian deal 59-61; 
Audiotronic proposition 78-80; 
CB radio 94-6; TV and video 
launch 100-1; BPI case 103-7; 
Young Businessman of the Year 
107; European build-up 140-4, 
150-2, 154-5, 158,260-8, 
301-2, 327; flotation 9-10, 
12-13, 15,80, 108;CPC464 
project 102, 116, 118-19, 122-7. 
129-31, 137; PCW8256 and Ant 
139, 160-5, 170-1, 173-7; other 
PC launches 222-4, 226-31, 
234, 316; Amsoft 164-6; share 
option scheme 256-8; Sinclair 
deal 189-99,201,202,206; 
Weinstock 4-5, 207-1 1, 293-4, 
338, 344-5; Murdoch and 
satellite TV 283-7, 290-304; 
City University lecture 244-9, 
251, 262-3; fight with EC 270-3 

Sugar, Ann (nee Simons): marriage 
and family 28, 32, 36-8, 43, 73, 
74, 109, 276; part of business 42, 
83-4, 98 

Sugar, Daniel (Alan's son) 277-8, 
299 



Sugar. Daphne (Alan's sister) 17, 
18. 32. 43. 58. 109 

Sugar. Derek (Alan's brother) 17. 
18. 20. 21, 27. 32,84, 275 

Sugar, Fay (Alan's mother); 
background and family 16-17, 
20-2, 28; Alan's successes 15, 
24, 25, 32,57-8, 109, 278 

Sugar, Nathan (Alan's father): 
background and family 16-17, 
18-22, 27, 28; cautious nature 
and relations with Alan 15, 
19-20, 29-30, 33, 35, 39, 84, 
109, 278 

Sugar, Shirley (Alan's sister) 17 

Sugar, Simon (Alan's son) 29, 43, 
227 

Sun, The 192, 299, 301 

Sunday Times 3, 227-8, 234, 262, 
274, 335, 346 

Taiwan: Amstrad's activities 7, 

61,82, 191,268,270,311-12, 

313 
Tandon 230 

Tarbuck, Jimmy 278, 280 
Templeman, Lord 106 
Texas Instruments 309, 31 1, 338 
Thatcher, Margaret, and 

Government 5, 7, 14, 205, 273, 

348 
Thorns, Peter 335 
Thomson, Lord 285 
Thomson, David 14, 111 
Thomson (of France) 131, 210, 

214-15,271,272,297 
Thorn, Sir Jules 217 
Thorn E.M.I. 189,214-15, 

217-18,351 



365 



DAVID THOMAS 



Time magazine 239, 250, 280 

Times, The 199 

Timex 189, 201 

Torch 138 

Toshiba 69, 70, 181, 247, 295, 311 

Tottenham site 83, 128, 150, 166 

Touche Ross 12, 91-2, 208-9, 214, 

254 
Tracey, Eric 254 

Unilever 59, 60, 88, 89, 255 
United Africa Company (UAC) 

59-60 
United States of America: 
electronic industry 117-18, 
135-6, 138, 222, 308-9, 342-3; 
Amstrad interests 143, 168, 222, 
261-4,273,328,333,335; 
Sugar's view of business 250-1 

Vannier, Marion 142-6, 151-4, 

156-9,257,261,327-8 
Video 263-4 
Virgin 285, 287, 330 

Wall Street Journal 138, 262, 280, 
311 

Wang Laboratories 342 

Waterman, Dennis 278 

Watkins, Bob: joins Amstrad 
85-7; first computer project 
115-17, 122, 126, 127, 129, 130, 
131; later projects 160-2, 168-9, 



170, 223-4; technical 

management 101, 127, 128, 172, 

199-200, 253-4, 269-70, 291-2; 

Sinclair deal 190, 191, 195-8; 

Spanish deal 149-51; SkyTV 

project 291-2, 294; share option 

257 
Weinstock, Arnold, Lord 4-5, 

121,207-8,291,292,343; 

Amstrad ventures 208-11, 

293-4, 338, 344-5 
Wesker, Arnold 20, 21-2 
West Germany: Amstrad interests 

140, 143, 222, 264-7, 273, 

301-2,313,328 
Western Digital 230, 316 
Which? 100, 169,227 
Which Computer? 176 
W'hitford, Mr Justice 104 
W.H.Smith 139, 141 
Wilding, Terry 185-6 
W T ilding Office Equipment 185-6, 

231 
Williams, Paul 11, 14, 110 
Wilson, Harold 30 
Wogan, Terry 112, 278 
Woolcock, Keith 233-5 
Woolworths 77, 78, 80, 97, 327 
Wozniak, Steve 135 

Young, Lord 272, 280-3, 331, 350 
Young, Barry 335, 341 

Zenith 343 



366 




David Thomas followed Amstrad closely while covering the 
electronics industry for the Financial Times. Born in 1954, he has 
written on a range of issues for the FT, including industry, 
employment, energy and education, having joined the paper after a 
spell as industrial correspondent on New Society. He was educated at 
Cambridge University, where he took a first in Social and Political 
Sciences and a Ph.D. in Philosophy. 



ISBN 7126 3518 I 
RANDOM CENTURY LTD 



ALAN 



'Probably Britain's greatest entrepreneur 9 
RUPERT MURDOCH 

'He's one of a new breed of British entrepreneurs. 

I would like to see people like that as role models 

for young people coming into business. I want 

people to say: damn it, if he can do it, I can.* 

LORD YOUNG, 

former Secretary of State for Trade and Industry 

'He has a talent which is not usually found in big 

electrical companies — his nose for the 

market. He has courage and nerve. He has 

commercial flair. He is young and independent. 

LORD WEINSTOCK, 

managing director of the General Electric Company 

'I thought Alan Sugar was an ideal person to take 

over my Spectrum computer business. He knew 

what was needed by the customer. He was very, 

very perceptive. And he knew where to get the 

product made efficiently.' 

SIR CLIVE SINCLAIR